Three Types of Server Hardware

Servers
A server is a computer or system that provides resources, data, services, or programs to other
computers, known as clients, over a network. In theory, whenever computers share resources with
client machines they are considered servers. There are many types of servers, including web
servers, mail servers, and virtual servers.
The term server can refer to the computer itself or to a program that delivers a service, such as
an email management system. Most hardware-related references concern the physical machine.

The server operating system (OS) is designed to process large workloads, deliver services and
support network-based operations. Common server OSes include Linux, Unix and Windows
Server.

Tower servers
Towers look very similar to desktop PCs. If you are not concerned about space you can house
more drives and more hardware into a single tower. However if you are looking at future expansion
and scalability, space can become a constraint with a tower server.A tower server is intended for
use as a server and is built in an upright standalone cabinet. The cabinet called a tower is similar
in dimensions to the cabinet of a tower-style desktop computer. Tower servers are good as they
provide easier cooling as the overall component density is comparatively low. The disadvantages
include the fact that a set of tower servers is bulkier and weigh more than a blade server or set of
rack servers. Also, the cabling for a big set of tower servers can be difficult and time-consuming.
Finally, many air-cooled tower servers all placed in the same location can be noisy since each
tower needs a dedicated fan.

Rack servers
Racks are designed to accommodate multiple servers, and are specially constructed to fit into
small spaces. They are usually stacked right on top of one other, which makes cooling the devices
difficult but is a big plus point when it comes to scalability. A rack server, also known as rack-
mounted server, is essentially a computer used as a server and is designed to be installed in a
framework called a rack. The rack constitutes of multiple mounting slots known as bays, which are
designed to secure a hardware unit held in place with screws. Unlike the tower server, the rack
server configuration simplifies cabling among network components. In an equipment rack filled
with servers, a special cooling system is necessary to prevent excessive heat build-up that would
otherwise occur when many power-dissipating components are confined in a small space.

Blade servers
Although blade servers are similar to rack servers in design, they are thinner, and also cost more
than the latter. A blade server is a server chassis which houses multiple thin, modular EC boards,
known as server blades. Each blade is actually a server, often dedicated to a single application.
The server blades are literally servers on a card that contain processors, memory, integrated
network controllers, and other input/output (IO) ports.

NOTE: If space is your main concern, you can go for racks. If space is not a concern but cooling
is then you can choose towers. And if money is not a concern and you want state-of-the-art server
hardware which is space saving and works efficiently go for blade servers.

TOP Servers Brands

There are several top server brands that can provide a business with the infrastructure they need
for their network. Dell, HP, Cisco, and Oracle all offer excellent hardware options, combining
superb reliability with the most current customization and software choices. HP, Dell, Oracle and
Cisco should be at the top of the shopping list when looking for hardware, whether new or
refurbished. All four have a reputation for producing quality equipment, though they are often
known for other technologies as well
HP’s Proliant line of hardware was developed by Compaq, back before it merged with HP. It is
consistently the best selling line of equipment every year, making it perhaps the most trusted line
of hardware on the market. Perhaps the most respected products in the Proliant line are the
DL360p G6, G7 or Gen8. They are popular choices for businesses looking for a refurbished option,
and feature Intel’s respected processor technology. The Gen8 is built with a robust Xeon E5-2600
processor, which can be fitted with anywhere from 4 to 12 processor cores. This makes it a
compact 1U piece of hardware that can be built with as much as 768 GB of RAM, giving it a lot of
punch for such a small package.
Dell’s PowerEdge is right there with HP’s Proliant line for top server brands. It offers superb 1U
and 2U options in the R410, R610, R710 and R810, and with the R420, R620, R720 and R820,
which were released in 2012. The R820 is one of the most robust pieces of equipment on the
market, offering as much as 1.5 TB of memory in a 2U form factor.
Cisco’s Unified Computing Systems (UCS) are relatively new to the market, but Cisco has plenty
of experience working with network infrastructure. Its M-series is designed to be as modular as
possible, primarily by disaggregating its secondary systems, like power and cooling.
Oracle hardware is known for its strong, built-in security systems and uses a proprietary version of
Linux known as Solaris, though some of its hardware is formatted for x86 applications. Its SPARC
line of hardware is the company’s premium line and can be scaled up to a TB of RAM and with
some of the best processors on the market. That makes Oracle’s equipment among the most
powerful technologies in the industry.

HP SERVERS
ProLiant is a brand of server computers that was originally developed and marketed by Compaq.
After Compaq merged with Hewlett-Packard (HP), HP retired its Netserver brand in favor of the
ProLiant brand.[1] The brand is currently marketed by Hewlett Packard Enterprise. HP ProLiant
systems led the x86 server market in terms of units and revenue during first quarter of
2010.[2] Proliant servers offers advanced capabilities such as dual power supplies with redundant
configuration support.

Modular Line (ML)

ML server models are tower-based. They aim towards maximum expandability.

Density Line (DL)

DL server models are rack-based. They aim towards a balance between density and computing
power.
Available models through product generations

Model Product generation

1 2 3 4 5 6 7 8 9 10

DL20 x x

DL60 x x

DL80 x x x

DL100 x x x

DL120 x x x x

DL140 x x x

DL145 x x x x

DL160 x x x x x[3]
Model Product generation

DL165 x x x x

DL180 x x x x x

DL185 x x x

DL320 x x x x x x x

DL325 x

DL360 x x x x x x x x x x[4]

DL365 x

DL370 x x x x x x x x

DL380 x x x x x x x x x x[5]

DL385 x x x x x x x

DL560 x x x x[6]

DL580 x x x x x x x x x

DL585 x x x x x

DL740 x

DL760 x x

DL785 x x
 Model Product generation

DL900 x

DL980 x x

DL1000 x

DL2000 x x x

Scalable Line (SL)

SL server models are rack-based. These models are mostly used in data centers and
environments where a maximum of computing power is desired.

Blade Line (BL)

BL server models are enclosure-based. They are made specially for use in a blade enclosure and
cannot be used without such. Blade systems aim towards maximum density and manageability at
limited rack [Link] are two models of blade enclosures: HPE BladeSystem c3000 Enclosure
(8 Bays for Blades), and HPE BladeSystem c7000 Enclosures (16 Bays for Blades).One
advantage of HP/HPE Blade Enclosures compared to competitors (such as IBM Blade Systems)
has been that the older generation enclosures have been able to accommodate new generation
BL servers just by upgrading the firmware for OA in the enclosure (Onboard Administrator).
However improvements to back-plane of the enclosure in the new generation enclosures have
enabled faster I/O capabilities (such as 10Gbit/s Ethernet adapters and switches, and Infiniband).
The physical design of the enclosures has not changed since the first versions (other than the
larger LCD screen at the front compared to the first generation enclosures, and new plastic
covers and HPE branding on the 3rd generation enclosures).

Details
ProLiant servers are separated into four main product lines - ML, DL, BL, SY, and Apollo - which
generally denote form factor. The ProLiant ML line comprises tower-based servers (convertible to
rack mount) with capacity for internal expansion of disks and interconnects, while the DL line
comprises general purpose rack mount servers. The BL line comprises blade servers which fit
within the HP BladeSystem, the SY comprises the Synergy Blades, and the Apollo line comprises
servers for scale out and High Performance Computing environments. The MicroServer product
line addresses small and home businesses.
ProLiant servers are also split into several series which denote processor configuration. The 100,
200, 300 and 400 series comprise single and dual socket capable systems, the 500 and 600 series
comprise quad socket capable systems, and the 700 and 900 series comprise eight socket capable
systems. The 900 series supports up to 80 CPU cores and up to 4 TB of RAM.
Models with a '0' in the last digit use Intel processors while models with a '5' in the last digit
use AMD processors.[7]
The ProLiant forms part of the HP Converged Systems, which use a common Converged
Infrastructure architecture for server, storage, and networking products. Designed to support 50 to
300 virtual machines, the HP ConvergedSystem 300 is configured with ProLiant servers. A system
administrator can manage ProLiant servers using HP OneView for converged infrastructure
management. HP also provide drivers a software for managing servers such as Management
Component Pack which includes hp online configuration utility (hponcfg), Agentless Management
Service amsd, Smart Storage Administrator (SSA) ssa, Smart Storage Administrator (SSA) CLI
ssacli and Smart Storage Administrator Diagnostic Utility (SSADU) CLI ssaducli.

Timeline
In February 2012, HP announced the ProLiant generation 8. In July 2013, HP announced a
new blade server-based ProLiant, the HP Moonshot Server.
DL580 Gen8 is a "middle generation" between Gen8 and Gen9. This server has some of new
features introduced in Gen9, primary there is available UEFI boot option.
Starting August 28, 2014, HP Proliant Gen9 series were available based on Intel Haswell chipset
and DDR4 memory. First were the HP ProLiant ML350 Gen9 Server and HP ProLiant BL460c
Gen9 Blade. Servers in this generation support both BIOS and UEFI.
Starting Q4 2017 HP Proliant Gen10 servers. were available. That year in November HPE
extended their Gen10 range to include AMD EPYC processors to further economics in server
virtualization using EPYC's impressive price v cores architectures claiming to lower cost per virtual
machine (VM) by 50%

HP Integrity Servers
HP Integrity is a series of server computers produced by Hewlett Packard
Enterprise (formerly Hewlett-Packard) since 2003, based on the Itanium processor. The Integrity
brand name was inherited by HP from Tandem Computers via Compaq.

In 2015 HP released the Superdome X line of Integrity Servers based on the x86
Architecture. It is a 'small' Box holding up to 8 dual Socket Blades and supporting up to 16
processors/240 cores (when populated with Intel Xeon E7-2890 or E7-2880 Processors).
Over the years HP had support for the Windows Server, HP-UX 11i, OpenVMS, NonStop, Red Hat
Enterprise Linux and SUSE Linux Enterprise Server operating systems on Integrity servers.[1]. As
of 2019 the two remain operating systems that are supported are HP-UX 11i and NonStop.

Entry-level servers
rx1600 series-The 1U rx1600 server is based on the zx1 chipset and has support for one or
two 1 GHz Deerfield Itanium 2 CPUs. The 1U rx1620 server is based on the zx1 chipset and has
support for one or two 1.3/1.6 GHz Fanwood Itanium 2 CPUs.

Common for the series is

 Support for up to 8 DIMMs

 Dual-channel Ultra-320 SCSI controller with support for two internal hard drives
 Support for two hot-swappable Ultra 320 drives
 One external Ultra 320 SCSI port
 10/100/1000BaseTX LAN port
 An additional 10/100BaseTX LAN port
 General-purpose RS-232 serial port
 Possibility for redundant power supplies
 Two USB ports
 Two PCI-X slots
Optional features are

 SCSI RAID controller

 Management processor card (required for support of Windows Server 2003)
 CD/DVD-ROM (required for installation of Windows Server 2003 and OpenVMS)

The series support five operating systems

 HP-UX 11i v2 or later

 OpenVMS for Itanium
 Windows Server 2003 for Itanium
 Windows Server 2008 for Itanium
 Linux with kernel that supports Itanium
rx2600 series-The 2U rx2600 server is based on the zx1 chipset and has support for one or
two 1.0/1.4 GHz Deerfield/Madison, 1.3 GHz Madison or 1.5 GHz Madison CPUs. The 2U rx2620
server is based on the zx1 chipset and has support for one or two 1.6 GHz Fanwood/Madison or
1.4/1.6 GHz Montecito CPUs. The 2U rx2660 server is based on the zx2 chipset and has support
for one or two 1.6 GHz Montvale or 1.42/1.66 GHz Montvale CPUs.

Common for the series is:

 Support for up to 12 DIMMs (8 DIMMs for the rx2660)

 Dual channel Ultra 320 SCSI controller with support for two internal hard drives
 Support for three hot-swappable Ultra 320 drives
 One external Ultra 320 SCSI port
 10/100/1000BaseTX LAN port
 Two general-purpose RS-232 serial ports
 Possibility for redundant power supplies
 Two USB ports
 Four PCI-X slots
Optional features are:

 SCSI RAID controller

 Management processor card (required for support of Windows Server 2003)
 CD/DVD-ROM (Required for installation of Windows Server 2003 and OpenVMS)
The series support four operating systems:

 HP-UX 11i v2 or later

 OpenVMS for Itanium
 Windows Server 2003 for Itanium
 Linux with kernel that supports Itanium

rx2800 series-The 2U rx2800 i2 server used the Intel Itanium 9300 series chipset with 8, 4 or
2 processor cores available. Among features common for the series:[3]

 Support for up to 24 DIMMs

 HP p410i 3 Gb SAS Controller
 Support for 8 hot-plug Serial Attached SCSI (SAS) small form factor (SFF) 2.5 inch drives
 Quad port 10/100/1000 Base TX LAN
 Six expansion slots
The series supports three operating systems: HP-UX 11i v3, OpenVMS v 8.4 for Itanium, and
Windows Server for Itanium.

rx3600-The 4U rx3600 is based on the zx2 chipset and has two CPU sockets which support
Montecito or Montvale processors. It supports up to 96 GB of main memory, using 24 four-gigabyte
DIMMs.

Standard features include:

 Support for up to 24 ECC memory PC2 4200 DDR2 DIMMs (with two optional 12-slot memory
carrier boards)
 Dual-port 10/100/1000Base TX LAN
 Integrated HP Integrated Lights-Out2 (iLO) management processor
 Choice of I/O backplane: PCI-X with 8 available PCI-X slots, or the PCI Express I/O backplane,
containing 4 x8 PCI Express slots, and 4 PCI-X slots. The system also has 2 reserved PCI-X
slots, one of which is used for a PCI-X dual-port 10/100/1000Base TX LAN card, and the other
is reserved for an optional internal Smart Array RAID card.
 N+1 hot-swappable fans
 Hot-swappable power supply
 Eight 2.5" Serial Attached SCSI (SAS) disk-drive slots
 8-port SAS host bus adapter, supporting two RAID 1 volumes and a hot spare
 Hot-swappable PCI-X I/O cards
Optional features:

 up to 8 SAS disk drives, with support for up to 72 GB 15K-RPM or 146 GB 10K-RPM disks.
 Redundant hot-swappable power supply
 8-port HP Smart Array P400 or P600, or 16-port P800 SAS RAID controllers available for
internal disks, supporting RAID 1/5/6 and hot spare disks.

rx4600 series-The 7U rx4610 utilizes Itanium 1 CPUs with support for up to four single-core
CPUs, 64GB RAM and 10 PCI 64-bit slots. The 4U rx4640 utilizes Itanium 2 CPUs with support
for up to four dual-core CPUs, 128 GB RAM and 6 PCI-X slots.
Both of these models come with integrated USB and video as default, which enables support for
Microsoft Windows operating systems. The rx4640 is architecturally the same box as the rp4440,
and can be changed from PA-RISC to Itanium 2 support with the flip of a switch.

rx5670-The discontinued 7U rx5670 server has four CPU sockets which support McKinley and
Madison processors. It is zx1-based and can have up to 48 DIMM slots, supporting 256 MB to
2 GB DIMMs which must be loaded in matched sets of four (quads). It has 9 PCI-X slots and 1 PCI
slot available.

rx6600-The 7U rx6600 is based on the zx2 chipset and has four CPU sockets that support
Montecito and Montvale CPUs. It supports 384 GB memory using 48 eight-gigabyte DIMMs.

Standard features include:

 Support for up to 48 ECC memory PC2 4200 DDR2 DIMMs (with the 48-slot memory board)
 Dual-port 10/100/1000Base TX LAN
 Integrated iLO2 management processor
 Choice of I/O backplane: PCI-X with 8 available PCI-X slots, or the PCI Express I/O backplane,
containing 4 x8 PCI Express slots, and 4 PCI-X slots. The system also has 2 reserved PCI-X
 slots, one of which is used for a PCI-X dual-port 10/100/1000Base TX LAN card, and the other
is reserved for an optional internal Smart Array RAID card.
 N+1 hot-swappable fans
 Hot-swappable power supply
 Sixteen 2.5" SAS disk drive slots
 8-port SAS host bus adapter, supporting two RAID 1 volumes and a hot spare
 Hot swappable PCI-X I/O cards
Optional features:

 Up to 16 SAS disk drives, with support for 72 GB 15K-RPM / 146 GB 10K-RPM or 300 GB
10K-RPM disks
 Redundant hot swappable power supply
 8-port HP Smart Array P400 or P600, or 16-port P800 SAS RAID controllers available for
internal disks, supporting RAID 1/5/6 and hot-spare disks

Mid-range servers
HP's mid-range and high-end (Superdome) servers are based on cell boards, sometimes
called cells,[4] which contain the chipset, Processors, memory, and I/O bus. This design allows the
servers to be divided into hardware partitions, or groups of cell boards. Each partition is able to
perform as if it were a separate server.

rx7600 series-The 10U rx7620 is based on the SX1000 chipset which supports both PA-
RISC and Itanium 2 CPUs. The 10U rx7640 is based on the SX2000 chipset which supports both
PA-RISC and Itanium 2 CPUs.

 Maximum of 2 cell boards

 4 CPU sockets per cell board
 16 DIMM slots per cell board
 Maximum of 4 SCSI disks and 2 tape and/or CD/DVD-ROM internally (half to each cell board)
 7 hot-pluggable I/O slots per cell board plus 1 core I/O slot per cell board for SX1000
The rx7600 series are the smallest cell-based servers from HP. Just like the bigger rx8600 (see
below) and the HP Superdome (see below), these servers can be partitioned, either as one big
partition (two cells in one partition) or as two independent cells.

rx8600 series-The 17U rx8620 is based on the SX1000 chipset which supports both PA-RISC
and Itanium 2 CPUs. The 17U rx8640 is based on the SX2000 chipset which supports both PA-
RISC and Itanium 2 CPUs.

 The rx8620 was announced in 2003, along with the rx4640 and rx7620[5]
 Maximum of four cell boards
 Four CPU sockets per cell board
 16 DIMM slots per cell board
 Maximum of four SCSI disks and two tape and/or CD/DVD-ROM internally (half to cell board
0, half to cell board 1)
 Eight hot-pluggable I/O-slots per cell board and 1 core I/O slot per cell board for SX1000
Just like the smaller rx7600 (see above) and the HP Superdome (see below), the rx8600 can be
partitioned using any combination of the four available cell boards (minimum of one, maximum of
four separated partitions).
The maximum number of partitions is four when used with an I/O-expander unit. Because each
partition requires an available I/O slot, and the rx8600 series' integrated I/O-chassis statically maps
its two I/O slots to cell board 0 and 1, an rx8600 series system is limited to two partitions unless
an IOX is installed.
Cells can be freely moved from a rx7600 series to a rx8600 series as long as the chipset is the
same on the cells, and the firmware is compatible.

High-end servers
The Superdome server is available in several models, including the SD-16, SD-32, and SD-64. HP
announced Superdome 2 in April 2010, offering resiliency improvements, a modular, bladed
design, common components and crossbar fabric that routes transactions to the optimal pathway
between blades and I/O. Superdome 2 addresses requirements for high-performance computing
by providing flexible scalability and fault tolerance necessary for mission-critical workloads.
In November 2011 HP announced Project Odyssey, a development roadmap to unify server
architectures on a single platform. The roadmap includes blades with Intel Xeon processors for the
HP Superdome 2 enclosure (code name “DragonHawk”) and the scalable c-Class blade
enclosures (code named “HydraLynx”), while supporting Windows and Linux environments with
features from HP-UX within the next two years.

DELL SERVERS
The Dell PowerEdge (PE) line is Dell's server computer product line.
Most PowerEdge servers use the x86 architecture. The early exceptions to this, the PowerEdge
3250, PowerEdge 7150, and PowerEdge 7250, used Intel's Itanium processor, but Dell
abandoned Itanium in 2005 after failing to find adoption in the marketplace. The partnership
between Intel and Dell remained close, with Intel remaining the exclusive source of processors in
Dell's servers until 2006. In May 2006 Dell announced that it also intended to develop servers
using AMD Opteron processors. The first Opteron-based PowerEdge systems, the PowerEdge
6950 and the PowerEdge SC1435, appeared in October 2006.

PowerEdge RAID Controller

Dell uses the name PowerEdge RAID Controller (PERC) for proprietary versions of
its RAID computer storage controllers. The related software in the PERC Fault Management Suite
offered facilities such as the Background Patrol read, which aims to fix bad sectors on online RAID
disks running under some of the PERC controllers around 2006. These cards were equipped with
hardware from LSI Corporation or Intel, 256 MBytes of memory (upgradeable on the 5/i to 512
MB), support up to 8x SATA 3.0 Gbit/s drives without the use of expanders. They had an optional
Battery Backup Unit (BBU) to allow more flexible use of the memory during writes, enhancing
performance in RAID5 and 6, and operate over the PCI Express interface

Chassis systems
Although PowerEdge is mainly used to refer to servers there are a few systems where the term
PowerEdge refers to systems of which servers are (just) a part. Examples of these usages are:
PowerEdge M1000e - the Dell blade-server system where the complete system uses the term
PowerEdge, and M1000e refers to the chassis and the complete combination of components in
them. The individual non-server components have also their own name in their 'own' family such
as PowerConnect M-switches or EqualLogic blade-SAN.
PowerEdge VRTX - the converged system consisting of (up to) 4 PowerEdge M-blade servers, the
built-in storage solution and the I/O networking module.

Model naming convention

Itanium servers
The Dell Itanium-based servers were introduced before this new naming-convention was
introduced and were only available as rack servers.

Since the introduction of the generation 10 servers in 2007 Dell adopted a standardized method
for naming their servers; the name of each server is represented by a letter followed by 3 digits.
The letter indicates the type of server: R (for Rack-mountable) indicates a 19" rack-mountable
server, M (for Modular) indicates a blade server, while T (for Tower) indicates a stand-alone
server.[15]
This letter is then followed by 3 digits:
The first digit refers to the number of CPU sockets in the system: 1 to 3 for one socket, 4 to 7 for
two sockets, and 9 for four sockets. 8 can be either two or four sockets depending on generation
and CPU maker
The second digit refers to the generation: 0 for Generation 10, 1 for Generation 11, and so on.
The third digit indicates the maker of the CPU: 0 for Intel or 5 for Advanced Micro Devices (AMD).
For example: The Dell PowerEdge M610 was a two-socket server of the 11th generation using an
Intel CPU while the R605 was a two-socket AMD-based rack-server of the 10th generation. This
letter is then followed by 3 digits:

 The first digit refers to the number of CPU sockets in the system: 1 to 3 for one
socket, 4 to 7 for two sockets, and 9 for four sockets. 8 can be either two or four sockets
depending on generation and CPU maker
 The second digit refers to the generation: 0 for Generation 10, 1 for Generation 11, and so
on.
 The third digit indicates the maker of the CPU: 0 for Intel or 5 for Advanced Micro
Devices (AMD).
For example: The Dell PowerEdge M610 was a two-socket server of the 11th generation using
an Intel CPU while the R605 was a two-socket AMD-based rack-server of the 10th generation.[18]
Prior to the Generation 10 servers, the naming convention was as follows:

 First digit – Height of the server in rack units

 Second digit – Generation of server (up to 9th generation)
 Third digit – Server type (5 for rack server, 0 for tower server, although tower servers could
be outfitted with a rack chassis)
 Fourth digit – Indicated whether blade or independent box (5 for blade, 0 for normal
independent box)
Example 1: PowerEdge 2650 ( 2 = 2U server, 6 = 6th generation, 5 = rack server, 0 = normal )
Example 2: PowerEdge 6950 ( 6 = 4U server, 9 = 9th generation, 5 = rack server, 0 = normal )
Example 3: PowerEdge 2800 ( 2 = [based on] 2U server 2850, 8 = 8th generation, 0 = tower
server, 0 = normal )
Example 4: PowerEdge 1855 ( 1 = 1U server, 8 = 8th generation, 5 = rack server, 5 = blade )
Most servers had a tower equivalent. For example, the PowerEdge 2800 was the tower
equivalent of the 2850. The naming applies to the tower version too, but the tower version will
usually be between 5U and 6U.

New naming conventions

Three digits
Since the introduction of the Generation 10 servers in 2007 Dell has adopted a standardized
method for naming their servers; the name of each server is now represented by a letter followed
by 3 digits. The letter indicates the type of server: R (for Rack-mountable) indicates a 19" rack-
mountable server, M (for Modular) indicates a blade server, whilst T (for Tower) indicates a
stand-alone server.
This letter is then followed by three digits.

 The first digit refers to the number of sockets in the system: 1 to 3 for one socket, 4 to 7 for
two sockets, and 8 or 9 for four sockets.
 The middle digit refers to the generation: 0 for Generation 10, 1 for Generation 11, and so
on.
 The third digit indicates the make of the CPU: 0 for Intel or 5 for AMD.
For example: The Dell PowerEdge M610 is a two-socket blade server of the 11th generation
using an Intel CPU whilst the R605 is a two-socket AMD-based rack-server of the 10th
generation.

Four digits
For four-digit naming convention:

 The first digit after the letter indicates the class of the system, with 1–5 defaulting to iDRAC
Basic and 6–9 defaulting to iDRAC Express.
 The second digit indicates the generation, with 0 for 10th generation, 1 for 11th generation
and so on.
 The third digit indicates the number of CPU sockets, 1 for one socket and 2 for two sockets.
 The fourth digit indicates the make of the CPU, 0 for Intel and 5 for AMD.
For example: The Dell PowerEdge R6415 model is a rack, mid-range, 14th generation, single
CPU socket system with AMD Processor.

Blade servers
Since Generation 10 there are models for the M1000e enclosure. The blade-servers in Generation
8 and Generation 9 are using another enclosure that is not compatible with the current M1000e
system. In form-factor there are two models: half-height and full-height. In an enclosure you can fit
8 full or 16 half-height blades (or a mix). Each server has two or four on-board NIC's and two
additional Mezzanine card-slots for additional I/O options: 1 Gb or 10 Gb Ethernet cards, Fibre
Channel HBA's or InfiniBand slots. Apart from USB connectors a blade-server doesn't offer direct
connections: all I/O goes via the midplane of the enclosure.
Early systems
Generation 1

Generation 2

Generation 3

Generation 4
Generation 5

Generation 6

Generation 7
Generation 8

Generation 9
Generation 10

Generation 11
Released by Dell in 2010
Generation 12
In March 2012 Dell introduced their 12th generation servers based on Intel Xeon. There are two
basic lines: 620 and 720.[124] On the 720 line, Dell currently offers two rack-model servers: the
Poweredge R720[125] and the R720XD[126] — where the latter offers the option to extend the system
to up to 26 internal disks.
The Poweredge 620 series offer models for rack, tower and a ½ height blade-server M620.[124] A ½
height blade means that you can fit up to 16 of those servers in one M1000e enclosure. The M520
and M620 can also be used in the new PowerEdge VRTX system. The new M420 is 1/4 height, so
32 fit in a M1000e chassis but does require a special full height holder that fits 4 M420's in one full-
height (=double) slot.
For the Generation 12 server-line the out of band server-management system iDRAC received a
new version: iDRAC 7. iDRAC allows you to access the server-console via a separate ethernet
connection allowing you to get access to the server even when there is no (working) operating
system or (normal) network connection available. It offers more or less the same functionality as a
network-enabled KVM switch, but with some additional options.
An overview of the offered servers as per July 2012[124]
Other 12th Generation servers, which are comparable in capabilities with the above detailed M or
R versions of the same model number, are:

 PE T320 Tower
 PE T420 Tower
Generation 13
In September 2014 Dell introduced their 13th generation servers based on Intel Xeon
Generation 14
In July 2017 Dell EMC introduced their 14th generation servers, adding support for latest Intel
Xeon Scalable Processors, better NVM support and other updates.[154]
PowerEdge SC Value Servers
Independent from the main generations of servers a value line was produced.

IBM SERVERS

IBM eServer was a family of computer servers from IBM Corporation. Announced in the year 2000,
it combined the various IBM server brands (AS/400, Netfinity, RS/6000, S/390) under one brand.
The various sub-brands were at the same time rebranded from:

 IBM AS/400 to IBM eServer iSeries, i for Integrated.

 IBM RS/6000 to IBM eServer pSeries, p for POWER
 IBM Netfinity to IBM eServer xSeries, x for eXtended architecture (with respect to
"commodity" Intel-based servers)
 IBM S/390 to IBM eServer zSeries, z for Zero downtime
In 2005 announced a new brand, IBM System as an umbrella for all IBM server and storage brands.
The name change was finished in 2006 when the IBM xSeries became the IBM System x then
Lenovo System x.

 IBM eServer iSeries became IBM System i

 IBM eServer pSeries became IBM System p
 IBM eServer xSeries became Lenovo System x
 IBM eServer zSeries became IBM Z
 IBM TotalStorage became IBM Storage
 IBM eServer BladeCenter became IBM BladeCenter
 IBM eServer 1350 became IBM System Cluster 1350
 IBM System i
The IBM System i is a line of midrange computer systems from IBM that use the IBM i operating
system. It was replaced by a modern standards-based (PCI, FiberChannel, RAID, etc) rack-friendly
server line, IBM Power Systems, in April 2008.
The platform was first introduced as the AS/400 (Application System/400) in June 1988, with the
operating system being called OS/400. In 2000, it was renamed to the eServer iSeries (with the
400 sometimes added). As part of IBM's re-branding initiative in 2006, it was again renamed
to System i, and the operating system was renamed to IBM i.
In April 2008, IBM announced its integration with the System p platform. The unified product line
is called IBM Power Systems and features support for the AIX and GNU/Linux operating systems
as well as for IBM i. Currently all Power Systems are manufactured in Guadalajara, Mexico, while
most of the engineering has remained in the U.S. where the combined operation included the
System p and POWER CPU teams.

Software
The IBM System i includes an extensive library-based operating system, IBM i, and is also capable
of supporting multiple instances of AIX, Linux, Lotus Domino, Microsoft Windows
2000 and Windows Server 2003. While IBM i, AIX, Linux and Lotus Domino are supported on
the POWER processors, Windows is supported with either single-processor internal blade servers
(IXS) or externally linked multiple-processor servers (IXA and iSCSI). iSCSI also provides support
for attachment of IBM Bladecenters. Windows, Linux, and VMware ESX(VI3) are supported on
iSCSI attached servers.

Hardware
CPU in AS/400, iSeries, i5, Power Systems[edit]
The System i5 uses POWER CPUs. These CPUs are developed and manufactured by IBM. The
POWER 4/5/5+ chips contain two cores. There are Multi-Chip Modules (MCM) available. They
have 2 CPUs (4 cores) or 4 CPUs (8 cores) in one MCM.

CPU Year Clock Speed Server-Models

since
Cobra (A10) 55 or 75 MHz 4xx, 5xx
1995

since
Muskie (A25/A30) 125 or 154 MHz 53x
1996

Apache (RS64) since

125 MHz 6xx, 150
(A35) 1997

NorthStar (RS64 since 200, 255 or

170, 250, 7xx, 650, S40, SB1[15]
II) 1998 262 MHz
 since
Pulsar (RS64 III) 450 MHz 270, 820
1999

since 400, 500, 540 or

IStar 820, 830, 840, SB2, SB3
2000 600 MHz

since 540, 600 or

SStar (RS64 IV) 270, 800, 810, 820, 830, 840
2000 750 MHz

since
POWER4 1.1 or 1.3 GHz 890
2001

since
POWER4+ 1.9 GHz 825, 870
2003

since
POWER5 1.5 or 1.9 GHz i5-520; i5-550; i5-570; i5-595
2004

1.5 GHz (2005)

since 1.9 GHz(2005) i5-520, i5-550, i5-515, i5-525
POWER5+
2005 2.2 GHz i5-570
2.3 GHz

3.5 GHz BladeCenter JS12, JS22

since
POWER6 4.2 GHz i5-570 (MMA)
2007
4.7 GHz M50, M25 & M15

3.6 GHz
3.8 GHz
since 4.0 GHz BladeCenter JS12, JS22, JS23, JS43
POWER6+
2009 4.2 GHz Power 520, 550, 560, 570, 575, 595
4.4 GHz
5.0 GHz

BladeCenter PS701, PS702, PS703,

PS704
3.3 GHz
PureSystem compute nodes p260, p460,
since 3.6 GHz
POWER7 p24L
2010 3.7 GHz
PowerLinux 7R1, 7R2
4.2 GHz
Power 710, 720, 730, 740, 750, 755, 760,
770, 780, 795
 BladeCenter PS701, PS702, PS703,
PS704
3.7 GHz PureSystem compute nodes p260, p460,
since
POWER7+ 4.2 GHz p24L
2012
4.4 GHz PowerLinux 7R1, 7R2
Power 710, 720, 730, 740, 750, 755, 760,
770, 780, 795

Power S812L, S814, S822, S822L, S824,

since
POWER8 2.5 GHz to 5.0 GHz S824L, S812LC, S821LC, S822LC
2014
Power E850, E870, E880

since Power AC922, L922, S914, S922, S924,

POWER9 4 GHz
2017 H922, H924, E950, E980

Models of AS/400, iSeries, i5 systems

Base
Model Year CPU Group
- CPW

1794 -
SB1, SB2, SB3 1997, 2000 P30, P40
16500

P05, P10, P20,

S10, S20, S30, S40 1997 45,4 - 4550
P30, P40, P50

1993, 1994,
P01, P02, P03 P05 7,3 - 16,8
1995

5500 -
MMA (9406) 2007 P30
76900

8150 -
MMA 2008 P30
76900

4800 -
M50 2008 P20
18000
 Base
Model Year CPU Group
- CPW

M25 2008 P10 4300 - 8300

M15 2008 P05 4300

JS43 2008

JS23 2008

JS22 2008 P10 13800

JS12 2008 P05 7100

F02, F04, F06, F10, F20, F25, F35, P05, P10, P20,
1993 5,5 - 177,4
F45, F50, F60, F70, F80, F90, F95, F97 P30, P40

E02, E04, E06, E10, E20, E25, E35, P10, P20, P30,
1992 4,5 - 116,6
E45, E50, E60, E70, E80, E90, E95 P40

D02, D04, D06, D10, D20, D25, D35,

1991 P10, P20, P30 3,8 - 56,6
D45, D50, D60, D70, D80

C04, C06, C10, C20, C25 1990 P10 3,1 - 6,1

B10, B20, B30, B35, B40, B45, B50,

1988, 1989 P10, P20 2,9 - 20
B60, B70

20000 -
890 2002 P50, P60
37400

7700 -
870 2002 P40, P50
20000

10000 -
840 2000,2001,2002 P40, P50
20200
 Base
Model Year CPU Group
- CPW

P20, P30, P40,

830 2000, 2002 1850 - 7350
P50

825 2003 P30 3600 - 6600

P05, P10, P20,

820 2000, 2001 100 - 3700
P30, P40

810 2003 P10, P20 750 - 2700

800 2003 P05, P10 300 - 950

740 1999 P40, P50 3660 - 4550

730 1999 P20, P30, P40 560 - 2890

720 1999 P10, P20, P30 240 - 1600

P05, P10, P20,

600, 620, 640, 650 1997 22,7 - 4550
P30, P40, P50

595 (9119) 2008 P60 294700

24500 -
595 2004–2006 P50, P60
216000

570 (9117) 2008 P30 104800

3300 -
570 2004–2006 P30, P40
58500

16700 -
570 2007 P40
58500
 Base
Model Year CPU Group
- CPW

3300 -
550 2004–2006 P20
14000

525 2007 P10 3800 - 7100

520 2004–2006 P05, P10, P20 500 - 7100

515 2007 P05 3800 - 7100

P10, P20, P30,

500, 50S, 510, 530, 53S 1995 18,7 - 650
P40

400, 40S, 436 1995 P05, P10 13,8 - 91,0

P10, P20, P30,

300, 30S, 310 1994 11,6 - 177,4
P40

270 2000 P05, P10, P20 50 - 2350

250 2000 P05 50 - 75

200, 20S, 236 1994 P05, P10 7,3 - 17,1

170 1998 P05, P10, P20, 30 - 1090

150 1996 P05 10,9 - 35,0

17,1 - 65,6
10S, 100, 135, 140 1995, 1993 P05, P10, P20
 IBM System p
The System p, formerly known as RS/6000, was IBM's RISC/UNIX-based server product
[Link] April 2008, IBM announced a rebranding of the System p and its unification with
the System i platform. The resulting product line

History
RS/6000 was originally a line of workstations and servers. The server line was then renamed to
the eServer pSeries in 2000 as part of its e-Server branding initiative.
In 2004, with the advent of the POWER5 processor, the family was rebranded the eServer p5. In
2005, following IBM's move to streamline its server and storage brands, worldwide, and
incorporating the System brand, with the Systems Agenda, the family was renamed yet again
as System p5.
System p generally uses the AIX operating system and, more recently, 64-bit versions of
the Linux operating system.

Models
BladeCenter[edit]
 IBM BladeCenter JS12 (POWER6)
 IBM BladeCenter JS22 (POWER6)
 IBM BladeCenter JS23 (POWER6)
 IBM BladeCenter JS43 (POWER6)

pSeries
 IBM eServer pSeries 610 (7028-6C1 & 6E1)
 IBM eServer pSeries 615 (7029-6C3, 7029-6E3) 1-2 POWER4 CPUs
 IBM eServer pSeries 620 (7025-F80, 6F0 & 6F1)
 IBM eServer pSeries 630 (7028-6C4, 7028-6E4) 1-4 POWER4 CPUs
 IBM eServer pSeries 640 (7026-B80) 1-4 POWER3-II CPUs
 IBM eServer pSeries 650 (7038-6M2) 2-8 POWER4 CPUs
 IBM eServer pSeries 655 (7039-651) 4-8 POWER4 CPUs
 IBM eServer pSeries 660 (7026-H80, 6H0, 6H1, M80 & 6M1)
 IBM eServer pSeries 670 (7040-671) 4-16 POWER4 CPUs
 IBM eServer pSeries 680 (7017 range)
 IBM eServer pSeries 690 (7040-681) 8-32 POWER4 CPUs
The IBM p690 was, at the time of its release in late 2001, the flagship of IBM's high
end Unix servers (pSeries) during the POWER4 era of processors. It was built to run IBM AIX Unix,
although it is possible to run a version of Linux minus some POWER4 specific features. It was
discontinued in late 2005.
It can support up to thirty-two 1.50, 1.70 or 1.90 GHz POWER4+ processors and 1 TB of RAM. It
weighs well over 1000 kg. It was used in a supercomputer at Forschungszentrum Jülich in 2004.

System p
System p was rebranded Power Systems in 2008.

 IBM System p 520 Express (POWER6)

 IBM System p 550 Express (POWER6)
 IBM System p 560 Express (POWER6)
 IBM System p 570 (POWER6)
 IBM System p 575 (POWER6)
 IBM System p 595 (POWER6)
 IBM System p5 505
 IBM System p5 505Q
 IBM System p5 510 (9110-510) 1-2 POWER5 CPUs
 IBM System p5 510 (9110-51A) 1-2 POWER5+ CPUs
 IBM System p5 510Q (9110-51A) 4 POWER5+ CPUs
 IBM System p5 520 (9111-520) 1-2 POWER5 CPUs
 IBM System p5 520Q
 IBM System p5 550 (9113-550) 1-4 POWER5 CPUs
 IBM System p5 550 Express(1.9G, 2.1G)
 IBM System p5 550Q
 IBM System p5 550Q Express (4-8 cores POWER5+ CPUs) (Model 9133-55A)
 IBM System p5 560Q
 IBM System p5 570 (9117-570) 2-16 POWER5 CPUs
 IBM System p5 575 (9118-575) 8-16 POWER5 CPUs
 IBM System p5 590 (9119-590) 8-32 POWER5 CPUs
 IBM System p5 595 (9119-595) 16-64 POWER5 CPUs

OpenPower

OpenPower was the name of a range of servers in the System p line from IBM. They featured
IBM's POWER5 CPUs and run only 64-bit versions of Linux. IBM's own UNIX variant, AIX is not
supported since the OpenPower servers are not licensed for this operating system.
There were two models available, with a variety of configurations.
Before 2005, OpenPower belonged to the eServer product line but were eventually rolled into the
IBM's Power Systems product portfolio.

 IBM eServer OpenPower 710 (9123-710) 1-2 POWER5 CPUs

 IBM eServer OpenPower 720 (9124-720) 1-4 POWER5 CPUs

IntelliStation POWER
 IBM IntelliStation POWER 265
 IBM IntelliStation POWER 275
 IBM IntelliStation POWER 185 (PowerPC 970)
 IBM IntelliStation POWER 285
BladeCenter[edit]
 IBM BladeCenter JS20 (PowerPC 970)
 IBM BladeCenter JS21 (PowerPC 970)

Lenovo System x
System x is a line of x86 servers produced by IBM – and later by Lenovo – as a sub-brand of
IBM's System brand, alongside IBM Power Systems, IBM System z and IBM System Storage. In
January 2014, IBM announced the sale of its x86 server business to Lenovo for $2.3 billion, in a
sale completed October 1, 2014.
History
Starting out as IBM PC Server, rebranded Netfinity, then eServer xSeries and now System x, these
servers are distinguished by being based on off-the-shelf x86 CPUs; IBM positions them as their
"low end" or "entry" offering. Previously IBM servers based on AMD Opteron CPUs did not share
the xSeries brand; instead they fell directly under the eServer umbrella. However, current AMD
Opteron-based servers fall under the System x brand.

IBM PC Server
PC Server range
 IBM PC Server 300
 IBM PC Server 310 (PCI/ISA)
 IBM PC Server 315 (PCI/ISA)
 IBM PC Server 320 (PCI/EISA)
 IBM PC Server 325 (PCI/EISA)
 IBM PC Server 330 (PCI/EISA)
 IBM PC Server 500 (MCA)
 IBM PC Server 520 (PCI/EISA or PCI/MCA)
 IBM PC Server 704 (PCI/EISA)
 IBM PC Server 720 (PCI/MCA)
Numbering scheme
 300 range for high-volume, entry level servers
 500 range for midrange
 700 range for high-end.

IBM Netfinity
Not to be confused with a different IBM product with a similar name, NetFinity (notice the capital
F).
Netfinity range
 IBM Netfinity 1000
 IBM Netfinity 3000, 3500
 IBM Netfinity 4000R, 4500R
 IBM Netfinity 5000, 5100, 5500, 5500-M10, 5500-M20, 5600
 IBM Netfinity 6000R
 IBM Netfinity 7000, 7000-M10, 7100, 7600
 IBM Netfinity 8500R
Numbering scheme
The numbering scheme started off similar to that of the IBM PC Servers, but additional ranges
were added, like the entry-level 1000 model later on. Models ending with an R, are rack-mount.

KVM Cabling scheme

Some Netfinity servers used IBM's C2T cabling scheme for Keyboard/Video/Mouse.

IBM eServer
IBM eServer range
IBM eServer was a marketing effort to put all of the diverse IBM server platforms under one header.
The AS/400 became the IBM eServer iSeries, the RS/6000 became the IBM eServer pSeries, the
S/390 mainframe became the IBM eServer zSeries and the Intel processor based IBM Netfinity
servers became the IBM eServer xSeries.
A few exceptions were however made

 IBM eServer 325, 326, 326m

 IBM eServer BladeCenter, BladeCenter T, BladeCenter H, BladeCenter HT
Numbering scheme
For marketing reasons the AMD processor based e325, e326 and e326m and the BladeCenter
which supports non-Intel processor products were not branded xSeries, but were instead placed
directly under the eServer brand. The xSeries brand was limited to only Intel-based server
products.
From a numbering perspective the AMD servers did fit into the xSeries range, under the similar
x335 and x336 Intel processor products. These numbers were not re-used in the xSeries range to
prevent confusion.

IBM eServer xSeries

xSeries range

 IBM eServer xSeries 100, 130, 135, 150

 IBM eServer xSeries 200, 205, 206, 206m, 220, 225, 226, 230, 232, 235, 236, 240, 250, 255,
260
 IBM eServer xSeries 300, 305, 306, 306m, 330, 335, 336, 340, 342, 345, 346, 350, 360, 365,
366, 370, 380, 382
 IBM eServer xSeries 440, 445, 450, 455, 460
Numbering scheme

 100 series are entry-level tower servers

 200 series are tower servers
 300 series are rack-mount servers
 400 series are rack-mount scalable servers
KVM Cabling scheme
Many xSeries servers used IBM's C2T cabling scheme for Keyboard/Video/Mouse.

IBM System x range

 IBM System x3105, x3100, x3100 M3, x3100 M4, x3100 M5
 IBM System x3200, x3200 M2, x3200 M3, x3250, x3250 M2, x3250 M3, x3250 M4, x3250
M5, x3250 M6
 IBM System x3300 M3, x3300 M4
 IBM System x3350
 IBM System x3400, x3400 M2, x3400 M3, x3450, x3455
 IBM System x3500, x3500 M2, x3500 M3, x3500 M4
 IBM System x3530 M3, x3530 M4
 IBM System x3550, x3550 M2, x3550 M3, x3550 M4, x3550 M5
 IBM System x3620 M3
 IBM System x3630 M3, x3630 M4
 IBM System x3650, x3650T, x3655, x3650 M2, x3650 M3, x3650 M4, x3650 M4 HD, x3650
M4 BD, 3650 M5
 IBM System x3690 X5
 IBM System x3750 M4
 IBM System x3755, x3755 M3
 IBM System x3800, x3850, x3850 M2, x3850 X5, x3850 X6
 IBM System x3950, x3950 M2, x3950 X5, x3950 X6

Lenovo System x range

These systems are effectively the same as the previous IBM branded models, but with a Lenovo
badge.

 Lenovo System x3100 M5

 Lenovo System x3250 M5, x3250 M6
 Lenovo System x3500 M5
 Lenovo System x3550 M4, x3550 M5
 Lenovo System x3650 M4, x3650 M5
 Lenovo System x3850 X6
 Lenovo System x3950 X6
 Lenovo NextScale
 Lenovo FlexSystem
Lenovo also had its own ThinkServer family of Intel servers. This family is technically less
advanced than System x. At the time of this writing, System x is being discontinued and replaced
by the Lenovo ThinkSystem family of Intel servers.

Enterprise eX5 architecture

Enterprise X4 architecture
Numbering scheme
2nd digit increments to show capability
3rd digit is a 0 for tower models, and 5 for rack-mount
4th digit is a 0 for Intel processors, and 5 for AMD Opteron.
Models with a T at the end are meant for Telco purposes.

IBM Z
IBM Z is a family name used by IBM for all of its z/Architecture mainframe computers from the
Z900 on. In July 2017, with another generation of products, the official family was changed to IBM
Z from IBM z Systems; the IBM Z family now includes the newest model the IBM z15, as well as
the z14 and the z13 (released under the IBM z Systems/IBM System z names), the IBM
zEnterprise models

Architecture
The zSeries, zEnterprise, System z and IBM Z families were named for their availability – z stands
for zero downtime. The systems are built with spare components capable of hot failovers to ensure
continuous operations.
Virtualization

Virtualization is required by default on IBM Z systems. First layer virtualization is provided by the
Processor Resource and System Manager (PR/SM) to deploy one or more Logical Partitions
(LPARs). Each LPAR supports a variety of operating systems. A hypervisor called z/VM can also
be run as the second layer virtualization in LPARs to create as many virtual machines (VMs) as
there are resources assigned to the LPARs to support them. The first layer of IBM Z virtualization
(PR/SM) allows a z machine to run a limited number of LPARs (up to 80 on the IBM z13).

List of models (reverse chronological order)

Since the move away from the System/390 name, a number of IBM Z models have been released.
These can be grouped into families with similar architectural characteristics.

IBM z15
 IBM z15 (8561) mainframe introduced on September 12, 2019
 IBM z15 Technical Introduction
 IBM z15 Technical Guide
IBM z14
 IBM z14 ZR1 (3907 series) single-frame mainframe introduced on April 10, 2018
 IBM z14 (3906 series) mainframe introduced on July 17, 2017
 Official IBM z14 mainframe product page
 IBM Redbooks z14 technical guide
IBM System z13
 z Systems z13s (2965 series), introduced on February 17, 2016
 z Systems z13 (2964 series), introduced on January 13, 2015
IBM System z10
The IBM System z10 servers supported more memory than previous generation systems and can
have up to 64 central processors (CPs) per frame. The full speed z10 processor's uniprocessor
performance was up to 62% faster than that of the z9 server, according to IBM's z10
announcement, and included these other features:
Specific models from this family include:

 z10 Business Class (2098 series), introduced on October 21, 2008

 z10 Enterprise Class (2097 series), introduced on February 26, 2008

IBM System z9
In July 2005, IBM announced a new family of servers – the System z9 family – with the IBM System
z9 Enterprise Class (z9 EC) and the IBM System z9 Business Class (z9 BC) servers. The System
z9 servers offered:
Specific models from this family include:

 z9 Business Class (2096 series), successor to the z890 and smallest z990 models (2006)
 z9 Enterprise Class (2094 series), introduced in 2005, initially as z9-109, beginning the
new System z9 line
IBM zSeries family
The zSeries family, which includes the z900, z800, z990 and z890, introduced IBM's newly
designed 64-bit z/Architecture to the mainframe world. The new servers provide more than four
times the performance of previous models. In its 64-bit mode the new CPU is freed from the 31-bit
addressing constraints of its predecessors. Major features of the eServer zSeries family:

 Based on z/Architecture (64-bit real and virtual addresses), as opposed to earlier ESA/390 (31-
bit) used in S/390 systems
 First zSeries Superscalar server (z990)
 Offers up to 32 central processors (CPs) per frame
 Frames can be coupled in up to a 32-frame Sysplex, with each frame physically separated up
to 100 kilometers
 Supports the z/OS, Linux on IBM Z, z/VM, z/VSE, and z/TPF operating systems
 Support of multiple I/O channel subsystem – or multiple Logical Channel Subsystem (LCSS).
The z990 allows for support of up to four LCSS – offering support for up to 4 times the
previous 256 channel limit

Specific models from this family included:

 z890 (2086 series), successor to the z800 and smaller z900 models (2004)
 z990 (2084 series), successor to larger z900 models (2003)
 z800 (2066 series), entry-level, less powerful variant of the z900 (2002)
 z900 (2064 series), for larger customers (2000)

Processor book
A processor book is a modular card in IBM mainframes that contains processors, memory, and I/O
connections. A multi-chip module is welded onto each processor book for the z196 model

IBM BladeCenter
The IBM BladeCenter was IBM's blade server architecture, until it was replaced by Flex System.
The x86 division was later sold to Lenovo in 2014

Versions
IBM BladeCenter (E)
The original IBM BladeCenter was later marketed as BladeCenter E[3] with 14 blade slots in 7U.
Power supplies have been upgraded through the life of the chassis from the original 1200 to 1400,
1800, 2000 and 2320 watt.
The BladeCenter (E) was co-developed by IBM and Intel and include

 14 blade slots in 7U
 Shared media tray with optical drive, floppy drive and USB 1.1 port
 One (upgradable to two) management modules
 Two (upgradable to four) power supplies
 Two redundant high-speed blowers
 Two slots for Gigabit Ethernet switches (can also have optical or copper pass-through)
 Two slots for optional switch or pass-through modules, can have additional Ethernet, Fibre
Channel, InfiniBand or Myrinet 2000 functions.
IBM BladeCenter T
BladeCenter T is the telecommunications company version[4] of the original IBM BladeCenter,
available with either AC or DC (48 V) power. Has 8 blade slots in 8U, but uses the same switches
and blades as the regular BladeCenter E. To keep NEBS Level 3 / ETSI compliant special Network
Equipment-Building System (NEBS) compliant blades are available

IBM BladeCenter H
Upgraded BladeCenter design with high-speed fabric options. Fits 14 blades in 9U. Backwards
compatible with older BladeCenter switches and blades.

Features

 14 blade slots in 9U
 Shared Media tray with Optical Drive and USB 2.0 port
 One (upgradable to two) Advanced Management Modules
 Two (upgradable to four) Power supplies
 Two redundant High-speed blowers
 Two slots for Gigabit Ethernet switches (can also have optical or copper pass-through)
 Two slots for optional switch or pass-through modules, can have additional Ethernet, Fibre
Channel, InfiniBand or Myrinet 2000 functions.
 Four slots for optional high-speed switches or pass-through modules, can have 10 Gbit
Ethernet or InfiniBand 4X.
 Optional Hard-wired serial port capability
IBM BladeCenter HT
Targets mid-sized customers by offering storage inside the BladeCenter chassis, so no separate
external storage needs to be purchased. It can also use 110 V power in the North American
market, so it can be used outside the datacenter. When running at 120 V , the total chassis
capacity is reduced.
Features

 6 blade slots in 7U
 Shared Media tray with Optical Drive and 2x USB 2.0 ports
 Up to 12 hot-swap 3.5" (or 24 2.5") SAS or SATA drives with RAID 0, 1 and 1E capability,
RAID 5 and SAN capabilities optional with two SAS RAID controllers
 Two optional Disk Storage Modules for HDDs, six 3.5-inch SAS/SATA drives each.
 4 hot-swap I/O switch module bays
 1 Advanced Management Module as standard (no option for secondary Management
Module)
 Two 950/1450-watt, hot-swap power modules and ability to have two optional 950/1450-watt
power modules, offering redundancy and power for robust configurations.
 Four hot-swap redundant blowers, plus one fan in each power supply.

Intel based
Modules based on x86 processors from Intel.
HS12
(2008) Features

 Celeron Single core 445 to Quad-Core Intel Xeon processors up to 2.83 GHz
 Six DIMM slots (up to 24 GB)
 Hot-swap drives
HS20
(2002–6) Features

 One or two Intel Xeon DP processors (single or dual-core)

 Four DIMM
 Option for one or two 2.5" drives (ATA100, SCSI U320 or Serial Attached SCSI (SAS)
depending on generation)
 Two Gigabit Ethernet ports
 One expansion slot for up to two additional ports (Fibre Channel storage, additional Ethernet,
Myrinet 2000 or InfiniBand)

HS21
(2007–8) This blade model can use the High-speed IO option of the BladeCenter H, but is
backwards compatible with the regular BladeCenter.
Features

 One or two Intel Xeon DP processors (dual or quad-core)

 Four DIMM slots
 Option for one or two SAS 2.5" drives
 Two Gigabit Ethernet ports
 One expansion slot for up to two additional ports (Fibre-Channel storage, additional Ethernet,
Myrinet 2000 or InfiniBand)
 One High-speed expansion slot for up to two additional ports (10 Gbit Ethernet or InfiniBand
4X)
HS21 XM
(2007–8) This blade model can use the High-speed IO option of the BladeCenter H, but is
backwards compatible with the regular BladeCenter.
Features

 One or two Intel Xeon DP processors (dual or quad-core)

 Eight DIMM slots
 Option for one SAS 2.5" drive or one or two SAS-based Solid State drives
 Two Gigabit Ethernet ports
 One expansion slot for up to two additional ports (Fibre-Channel storage, additional
Ethernet, Myrinet 2000 or InfiniBand)
 One High-speed expansion slot for up to two additional ports (10 Gbit Ethernet
or InfiniBand 4X as well as additional Fibre-Channel or Ethernet ports)
HS22
(2009–11) Features

 One or two Intel Xeon 5500 or 5600 series processors (up to 3.6 GHz 4c or 3.46 GHz 6c)
 Twelve DDR3 VLP DIMM slots (up to 192 GB of total memory capacity)
 Option for two hot swap SAS 2.5" drives or Solid State drives (RAID 0 and 1 are possible)
 Two Gigabit Ethernet ports (Broadcom 5709S)
 1 CIOv slot (standard PCI-Express daughter card) and 1 CFFh slot (high-speed PCI-Express
daughter card) for a total of 8 ports of I/O to each blade, including 4 ports of high-speed I/O
 Requires Advanced Management Module
HS22v
(2010–11) Features are very similar to HS22 but:

 18 DDR-3 VLP DIMM slots, up to 288 GB of total memory capacity

 up to two 1.8" disks (SSD, not hot swapable)
 Requires Advanced Management Module
HS23
(2012) Features

 Single-wide
 One or Two Intel Xeon E5-2600
 16 Dimm slots, up to 1600mhz
 2 hot-swappable HDDs (SATA/SAS) or SSDs
 Dual 10G/1G Ethernet onboard expandable to 4x10
 Virtual Fabric / vNic's onboard
 Requires Advanced Management Module
HS23E
(2012) Features:

 Single-wide
 One or Two Intel Xeon E5-2400
 12 Dimm slots, up to 1600mhz
 2 hot-swappable HDDs (SATA/SAS) or SSDs
 Dual Gigabit Ethernet onboard ports with TOE
 Requires Advanced Management Module
HS40
(2004) Features

 Double-wide (needs 2 slots)

 One to four Intel Xeon MP processors
 Eight DIMM slots
 Option for one or two ATA100 2.5" drives
 Four Gigabit Ethernet ports
 Two expansion slots for up to four additional ports (Fibre Channel storage, additional
Ethernet, Myrinet 2000 or InfiniBand)

HC10
(2008) This blade model is targeted to the workstation market
Features

 Single Intel Core 2 Duo processor

 Four DIMM slots (8 GB Max)
 NVidia video adapter
 One SATA 60 GB HDD
 Two Gigabit Ethernet ports
HX5
(2010–11) This blade model is targeted at the server virtualization market.
Features

 2 expandable to 4 Intel E7, 6500 or 7500 series Xeon (4-10 cores per CPU, up to 2.67 GHz)
 Sixteen DIMM slots (256 GB Max). Expandable to 40 slots with a 24 DIMM MAX 5 memory
blade (640GB total).
 Two Gigabit Ethernet ports per blade.
AMD based
Modules based on x86 processors from AMD.

LS20
(2005-6) Features

 One or two AMD Opteron processors (single or dual-core)

 Four DIMM slots for DDR1 VLP DIMM's
 Option for one or two SCSI U320 2.5" drives
 Two Gigabit Ethernet ports
 One expansion slot for up to two additional ports (Fibre Channel storage, additional
Ethernet, Myrinet 2000 or InfiniBand)
LS21
(2006)
This blade model can use the high-speed I/O of the BladeCenter H, but is also backwards
compatible with the regular BladeCenter.
Features:

 One or two AMD Opteron processors (dual-core), support 65nm Quad core after BIOS
update (tested)
 Eight DIMM slots for up to 32 GB of RAM
 Option for one SAS or SATA 2.5" drive
 Two Gigabit Ethernet ports
 One expansion slot for up to two additional ports (Fibre-Channel storage, additional
Ethernet, Myrinet 2000 or InfiniBand)
 One High-speed expansion slot for up to two additional ports (10 Gbit Ethernet
or InfiniBand 4X)
LS22
(2008) Upgraded model of LS21.
Features:

 One or two 45nm AMD Opteron processors (quad or hex-core)

 Eight DIMM slots for up to 64 GB of RAM
 Option for two SAS or SATA 2.5" drive
 Two Gigabit Ethernet ports
 One expansion slot for up to two additional ports (Fibre-Channel storage, additional
Ethernet, Myrinet 2000 or InfiniBand)
 One High-speed expansion slot for up to two additional ports (10 Gbit Ethernet
or InfiniBand 4X)
LS41
(2006–7) This blade model can use the High-speed IO option of the BladeCenter H, but is
backwards compatible with the regular BladeCenter
Features

 Double-wide (needs 2 slots)

 One to four AMD Opteron processors (dual-core)
 Sixteen DIMM slots for up to 64 GB of RAM
 Option for one or two SAS 2.5" drives
 Four Gigabit Ethernet ports
 Two expansion slots for up to four additional ports (Fibre-Channel storage, additional
Ethernet, Myrinet 2000 or InfiniBand)
 One High-speed expansion slot for up to two additional ports (10 Gbit Ethernet
or InfiniBand 4X)
 Data Storage Capacity ?

LS42
(2008–9) Upgraded model of LS41.
Features

 Double-wide (needs 2 slots)

 One to four AMD Opteron processors (quad or hex-core)
 Sixteen DIMM slots for up to 128 GB of RAM
 Option for one or two SAS 2.5" drives
 Four Gigabit Ethernet ports
 Two expansion slots for up to four additional ports (Fibre-Channel storage, additional
Ethernet, Myrinet 2000 or InfiniBand)
 One High-speed expansion slot for up to two additional ports (10 Gbit Ethernet
or InfiniBand 4X)

Power based
Modules based on PowerPC- or Power ISA-based processors from IBM.

JS20
(2006) Features

 Can run AIX or Linux

 Two PowerPC 970 processors at 1.6 or 2.2 GHz
 Four DIMM slots for PC2700 ECC memory (max 8 GB)
 Option for one or two ATA100 2.5" drives
 Two Gigabit Ethernet ports
 One expansion slot for up to two additional ports (Fibre Channel storage, additional
Ethernet, Myrinet 2000 or InfiniBand)
JS21
(2006) This blade model can have the High-speed IO option of the BladeCenter H, but is
backwards compatible with the regular BladeCenter
Features

 Can run AIX or Linux

 Can do virtualization, offering Dynamic Logical Partitioning (DLPAR) capabilities.
 Two PowerPC 970FX single-core processors at 2.7 GHz or two PowerPC 970MP dual-core
processors at 2.5 GHz
 Four DIMM slots for PC2-3200 or PC2-4200 ECC memory (max 16 GB)
 Option for one or two SAS 2.5" drives
 Two Gigabit Ethernet ports
 One expansion slot for up to two additional ports (Fibre Channel storage, additional Ethernet,
Myrinet 2000 or InfiniBand)
 One High-speed expansion slot for up to two additional ports (10 Gbit Ethernet
or InfiniBand 4X)
JS22
(2009) Features

 Can run IBM i, AIX or Linux

 Can do virtualization, offering Dynamic Logical Partitioning (DLPAR) capabilities.
 Two POWER6 dual-core processors at 4.0 GHz
 Four DIMM slots ECC Chipkill DDR2 SDRAM memory (max 32 GB)
 One 2.5" SAS drive up to 146 GB
 Integrated Virtualization Manager(IVM)
 Two Gigabit Ethernet ports card
 One expansion slot for up to two additional ports (Fibre Channel storage, additional
Ethernet, Myrinet 2000 or InfiniBand)
 One High-speed expansion slot for up to two additional ports (10 Gbit Ethernet
or InfiniBand 4X)
JS23
(2009) Features

 Can run IBM i, AIX or Linux

 Can do virtualization, offering Dynamic Logical Partitioning (DLPAR) capabilities.
 Two POWER6 dual-core processors at 4.2 GHz
 64 MB L3 cache (32 per processor)
 Four DIMM slots ECC Chipkill DDR2 SDRAM memory (max 64 GB)
 One 2.5" SAS up to 300 GB or one 69 GB SSD drive
 Integrated Virtualization Manager(IVM)
 Two Gigabit Ethernet ports card
 One expansion slot for up to two additional ports (Fibre Channel storage, additional
Ethernet, Myrinet 2000 or InfiniBand)
 One High-speed expansion slot for up to two additional ports (10 Gbit Ethernet
or InfiniBand 4X)
 One CIOv PCIe expansion slot

JS43 Express
Features

 Can run IBM i, AIX or Linux

 Can do virtualization, offering Dynamic Logical Partitioning (DLPAR) capabilities.
 Four POWER6 dual-core processors at 4.2 GHz
 128 MB L3 cache (32 per processor)
 Eight DIMM slots ECC Chipkill DDR2 SDRAM memory (max 128 GB)
 Zero to two 2.5" SAS up to 300 GB or zero to two 69 GB SSD drives
 Integrated Virtualization Manager(IVM)
 Two Gigabit Ethernet ports card
 One expansion slot for up to two additional ports (Fibre Channel storage, additional
Ethernet, Myrinet 2000 or InfiniBand)
 One High-speed expansion slot for up to two additional ports (10 Gbit Ethernet
or InfiniBand 4X)
 Two PCIe CIOv expansion slots
 Double-wide form factor
JS12 Express
Features
 Can run IBM i, AIX or Linux
 Can do virtualization, offering Dynamic Logical Partitioning (DLPAR) capabilities.
 One POWER6 dual-core processor at 3.8 GHz
 Eight DIMM slots ECC Chipkill DDR2 SDRAM memory (max 64 GB)
 Zero to two 2.5" SAS drive up to 146 GB
 Integrated Virtualization Manager(IVM)
 Two Gigabit Ethernet ports, with optionan dual Gbit card
 One expansion slot for up to two additional ports (Fibre-Channel storage, additional
Ethernet, Myrinet 2000 or InfiniBand)
 One High-speed expansion slot for up to two additional ports (4 Gbit/s Fibre Channel, iSCSI,
or InfiniBand 4X)
PS700
Features

 Can run IBM i, AIX or Linux on Power

 Can do virtualization, offering Dynamic Logical Partitioning (DLPAR) capabilities.
 One POWER7 quad-core processor at 3 GHz
 max 64 GB
 Two Gigabit Ethernet ports, with optionan dual Gbit card
PS701
Features are very similar to PS700, but

 Eight-core processor
 max 128 GB
PS702
Think two PS701 tied together back-to-back, forming a double-wide blade

PS703
Features are very similar to PS701, but

 Two eight-core processors at 2.4 GHz

PS704
Think two PS703 tied together back-to-back, forming a double-wide blade.

Cell based
Modules based on Cell processors from IBM.

QS20
Features

 Double-wide (needs 2 slots)

 Can run Linux
 Two Cell processors at 3.2 GHz
 1 GB XDRAM Memory (512 MB per processor)
 40 GB IDE100 2.5" drive
 Two Gigabit Ethernet ports
 Optional InfiniBand 4X connectivity
QS21
Features

 Single-wide
 Can run Linux
 Two Cell processors at 3.2 GHz
 2 GB Memory (1 GB per processor)
 Two Gigabit Ethernet ports
 One expansion slot for up to two additional ports (Fibre-Channel storage, additional
Ethernet, Myrinet 2000 or InfiniBand)
 One High-speed expansion slot for up to two additional ports (10 Gbit Ethernet
or InfiniBand 4X)
QS22
Features

 Single-wide
 Can run Linux
 Two PowerXCell 8i processors at 3.2 GHz
 Up to 32 GB of DDR2 SDRAM RAM memory
 Two Gigabit Ethernet ports
 Expansion slots for SAS daughter card, InfiniBand 4X DDR daughter card and 8 GB uFDM
Flash Drive

Sun servers

SPARC Enterprise
The SPARC Enterprise series is a range of UNIX server computers based on the SPARC
V9 architecture. It was co-developed by Sun Microsystems and Fujitsu, and introduced in 2007.
They were marketed and sold by Sun Microsystems (later Oracle Corporation, after
their acquisition of Sun), Fujitsu, and Fujitsu Siemens Computers under the common brand of
"SPARC Enterprise", superseding Sun's Sun Fire and Fujitsu's PRIMEPOWER server product
lines.
Since 2010, servers based on new SPARC CMT processors (SPARC T3 and later) have been
branded as Oracle's SPARC T-Series servers.
Model range

SPARC64 processor based models (M-series)

The midrange and high-end SPARC64 VI, SPARC64 VII, SPARC64 VII+ processor based servers
are designated "M-series". The "M" indicates RAS features similar to mainframe class machines.

 M3000 - 1 processor socket, 2U rack-mount

 M4000 - Up to 4 processor sockets, 6U rack-mount
 M5000 - Up to 8 processor sockets, 10U rack-mount
 M8000 - Up to 16 processor sockets, one data center rack
 M9000 - Up to 64 processor sockets, one or two data center racks
Processor Types
The SPARC64 VI is a dual-core processor, with each core featuring two-way vertical multi-
threading (VMT).[2] A M9000 server configured with the maximum number of processors supports
running 256 concurrent threads. VMT is a coarse-grained multi-threading implementation. Each
core in the SPARC64 VI can handle two strands or threads. VMT switches execution from one
strand to the other on the basis of events. To execute instructions from another thread, the pipeline
must be saved/flushed and switched to the registers for the other thread. These events include L2
cache misses, a hardware timer exception, interrupts, or some multi-threading-related control
instructions. This is also called Switch On Event (SOE) threading.
In 2008, Fujitsu released the SPARC64 VII, a quad-core processor, with each core featuring two-
way simultaneous multi-threading. Existing M-class servers will be able to upgrade to the
SPARC64 VII processors in the field.[3]
In 2010, Fujitsu released the SPARC64 VII+,[4] running at higher frequency and with a larger L2
cache than its predecessor. A SPARC64 VII or SPARC64 VII+ processor module includes four
physical cores, where each core can execute two threads. Each physical core is able to run both
threads simultaneously. With SMT, there is no context-switch time and the two threads share the
instruction pipeline smoothly. When both are ready to run, they alternate cycles
for superscalar instruction issue, and share the functional units according to need.
An important capability of the M-Series is the ability to mix processor generations and clock speeds
in the same system and domain. All M-Series servers can have both SPARC64 VI and SPARC64
VII CPUs installed and they will run at their native speeds, with no clocking down to the slowest
CPU.
UltraSPARC T processor based models (T-series)
The UltraSPARC T1-based Sun Fire T1000 and T2000 are rebranded to SPARC Enterprise:

 T1000 - 1 processor socket, 1U rack-mount

 T2000 - 1 processor socket, 2U rack-mount
In October 2007, Sun added the UltraSPARC T2-based servers to the SPARC Enterprise line:

 T5120 - 1 processor socket, 1U rack-mount

 T5220 - 1 processor socket, 2U rack-mount
In April 2008, Sun added the UltraSPARC T2 Plus-based servers to the SPARC Enterprise line:

 T5140 - 2 processor sockets, 1U rack-mount

 T5240 - 2 processor sockets, 2U rack-mount
In October 2008, Sun released the 4-way SMP UltraSPARC T2 Plus-based server:

 T5440 - 4 processor sockets, 4U rack-mount

Operating systems
SPARC Enterprise models are licensed to run Solaris 10 and Solaris 11.

Partitioning and virtualization

The M-series supports Dynamic Domains and Dynamic Reconfiguration,which enable a single
machine to be divided into multiple electrically isolated partitions.
The UltraSPARC T1, UltraSPARC T2/T2+ models partition the system using Logical Domains.
Both M-series and T-series models support Solaris Containers, which supports a maximum of 8191
non-global zones in each Dynamic Domain or Logical Domain.

SPARC T series
The SPARC T-series family of RISC processors and server computers, based on the SPARC
V9 architecture, was originally developed by Sun Microsystems, and later by Oracle
Corporation after its acquisition of Sun. Its distinguishing feature from earlier SPARC iterations is
the introduction of chip multithreading (CMT) technology, a multithreading, multicore design
intended to drive greater processor utilization at lower power consumption.
The first generation T-series processor, the UltraSPARC T1, and servers based on it, were
announced in December 2005. As later generations were introduced, the term "T series" was used
to refer to the entire family of processors.[

Pre-Oracle era
Sun Microsystems' Sun Fire and SPARC Enterprise product lines were based on early generations
of CMT technology. The UltraSPARC T1 based Sun Fire T2000 and T1000 servers were launched
in December 2005 and early 2006, [Link] were later rebranded to match the name of
the UltraSPARC T2 and T2 Plus based Sun SPARC Enterprise T5**0 servers

SPARC T3
In September 2010, Oracle announced a range of SPARC T3 processor based servers. These are
branded as the "SPARC T3" series, the "SPARC Enterprise" brand being dropped.
The SPARC T3-series servers include the T3-1B, a blade server module that fits into the Sun
Blade 6000 system. All other T3 based servers are rack mounted systems. Subsequent T-series
server generations also include a blade server in the same Sun Blade 6000 form factor.

SPARC T4
On September 26, 2011, Oracle announced a range of SPARC T4-based [Link] systems
use the same chassis as the earlier T3 based systems. Their main features are very similar, with
the exception of:

 T4 CPU instead of T3 CPU, with complete core redesign

 doubled RAM capacity
 small changes in mass storage capacity

SPARC T5
On March 26, 2013, Oracle announced refreshed SPARC servers based on the new SPARC
T5 microprocessor, which the company claims is "the world's fastest". In the T5 range of servers,
the single socket rackmount server design was deprecated, while a new eight-socket rackmount
server was introduced.

SPARC M7
On October 26, 2015, Oracle announced a family of systems built on the 32-core, 256-thread
SPARC M7 microprocessor. Unlike prior generations, both T- and M-series systems were
introduced using the same processor.
Partitioning and virtualization
SPARC T-series servers can be partitioned using Oracle's Logical Domains technology. Additional
virtualization is provided by Oracle Solaris Zones (aka Solaris Containers) to create isolated virtual
servers within a single operating system instance. Logical Domains and Solaris Zones can be used
together to increase server [Link]
SPARCstation
The SPARCstation, SPARCserver and SPARCcenter product lines were a series of SPARC-
based computer workstations and servers in desktop, desk side (pedestal) and rack-based form
factor configurations, developed and sold by Sun Microsystems.

Models
Desktop and deskside SPARCstations and SPARCservers of the same model number were
essentially identical systems, the only difference being that systems designated as servers were
usually "headless" (that is, configured without a graphics card and monitor), and were sold with a
"server" rather than a "desktop" OS license. Most desktop SPARCstations and SPARCservers
shipped in either "pizzabox" or "lunchbox" enclosures, a significant departure from earlier Sun and
competing systems of the time. The SPARCstation 1, 2, 4, 5, 10 and 20 were "pizzabox" machines.

Pizzabox" systems

"Lunchbox" systems
Integrated monitor/portable systems

Server systems

Sun Enterprise
Sun Enterprise is a range of UNIX server computers produced by Sun Microsystems from 1996
to 2001hese systems are based on the 64-bit UltraSPARC microprocessor architecture and related
to the contemporary Ultra series of computer workstations. Like the Ultra series, they run Solaris.
Various models, from single-processor entry-level servers to large high-end multiprocessor servers
were produced. The Enterprise brand was phased out in favor of the Sun Fire model line from
2001 onwards.

Ultra workstation-derived servers

The first UltraSPARC-I-based servers produced by Sun, launched in 1995, are the UltraServer
1 and UltraServer 2. These are server configurations of the Ultra 1 and Ultra 2 workstations
respectively. These were later renamed Ultra Enterprise 1 and Ultra Enterprise 2 for consistency
with other server models. Later these were joined by the Ultra Enterprise 150, which comprises
an Ultra 1 motherboard in a tower-style enclosure with 12 internal disk bays.
In 1998, Sun launched server configurations of the UltraSPARC-IIi-based Ultra 5 and Ultra 10
workstations, called the Enterprise Ultra 5S and Enterprise Ultra 10S respectively.

Entry-level servers
The Sun Enterprise 450 is a rack-mountable entry-level multiprocessor server launched in 1997,
capable of up to four UltraSPARC II processors. The Sun Enterprise 250 is a two-processor
version launched in 1998. These were later joined by the Enterprise 220R and Enterprise
420R rack-mount servers in 1999. The 220R and 420R models are respectively based on the
motherboards of the Ultra 60 and Ultra 80 workstations. The 250 was replaced by the Sun
Fire V250, the 450 by the Sun Fire V880. The 220R was superseded by the Sun Fire 280R and
the 420R by the Sun Fire V480.
Ultra Enterprise X000/Enterprise X500 mid-range
servers
In 1996, Sun replaced the SPARCserver 1000E and SPARCcenter 2000E models with the Ultra
Enterprise 3000, 4000, 5000 and 6000 servers. These are multiprocessor servers based on a
common hardware architecture incorporating the Gigaplane packet-switched processor/memory
bus and UltraSPARC-I or II processors.
The 3000 model is a deskside server configurable with up to six processors and 10 internal disks,
while the 4000 is a rack-mount system with up to 14 processors. The 5000 is essentially a 4000 in
a rack cabinet and the 6000 is a cabinet-housed data center server with up to 30 processors. In
1999, the Enterprise 3500, 4500, 5500 and 6500 models were announced. These are upgraded
X000 systems, with a faster Gigaplane busThe Enterprise X500 series were replaced by the Sun
Fire 3800/4800/6800 line.

Enterprise 10000
The Enterprise 10000, E10k or Starfire is a high-end multiprocessor data center server capable
of being configured with up to 64 UltraSPARC II processors. The Starfire is based around the fault-
tolerant Gigaplane-XB processor/memory interconnect. Like the X000 and X500 series servers,
the Starfire incorporates many high-availability features, including the ability to be partitioned into
multiple "domains", each of which can be booted individually to run its own instance of Solaris. It
is also possible to remove resources from a running domain with short notice and reassign freed
resources to other domains. Domain granularity is one CPU board (single system may have 1-16
of them). A single CPU board can carry up to 4 processors, 4GB of RAM and 4 SBUS IO boards.
The Starfire contains one or two controller modules which are connected via Ethernet to an
external computer, the System Service Processor (SSP). The controller modules interface with the
system "centerplane" via JTAG and control the partitioning of available CPUs, memory and I/O
devices into one or more domains, each of which is in effect a distinct computer. The Starfire was
superseded by the Sun Fire 12K/15K models.

Sun Fire 15K

The Sun Fire 15K (codenamed Starcat) was an enterprise-class server computer from Sun
Microsystems based on the SPARC V9 processor architecture. It was announced on September
25, 2001 in New York City, superseding the Sun Enterprise 10000.
The Sun Fire 15K supported up to 106 UltraSPARC III processors (up to 1.2 GHz), or
72 UltraSPARC IVs (up to 1.35 GHz) installed across 18 system boards (Uniboards, containing
CPU sockets and RAM slots). With the UltraSPARC III, Sun supported up to 17 dual-socket
"MaxCPU" processor cards in place of I/O mezzanine cards, a configuration not supported with
UltraSPARC IV. Maximum physical RAM per system is 576 GB. A maximum of 72 PCI I/O slots
are available.

The system can be divided into a maximum of 18 secure independent domains, each of which is
a separate machine with its own filesystems, root password and the ability to run different versions
of Solaris.

Sun Fire 12K

The Sun Fire 12K (codenamed Starkitty), was a reduced configuration version of the 15K,
introduced in April 2002. It supported a maximum of 52 processors, and was intended to fill a
position in Sun's server product line between the 15K and the Sun Fire 6800.
Sun Fire E25K
The Sun Fire E25K was announced in February 2004. Its base cabinet is identical to the 15K, with
the only difference between the two systems being the processor boards installed. It reached end-
of-life in January, 2009, and was superseded by the Sun SPARC Enterprise M9000 server.
The E25K supports up to 72 dual-core UltraSPARC IV+ processors (up to 1.95 GHz). As with
UltraSPARC IV-based 15K systems, the "MaxCPU" option was not offered for E25K systems.
Overall system bandwidth is claimed as up to 172.8 GB/s aggregate, up to 115.2 GB/s peak, and
up to 43.2 GB/s sustained. For overall I/O bandwidth, up to 35.8-GB/s sustained. Up to 64 GB of
RAM per board is possible with a maximum of 1.15 TB of RAM for a single domain. Up to 72 hot
swappable PCI-X I/O slots; 54 slots are 90 MHz, 18 slots are 33 MHz. It also supports 10/100
BaseT Ethernet, Gigabit Ethernet, UltraSCSI (LVD and HVD), ATM, FC-AL, HSI and SCI.

Sun Fire T2000

The Sun Fire T2000 server is a system engineered by Sun Microsystems for applications including
Web 2.0 and databasing.

Hardware
The T2000 is powered by a four-core, six-core or eight-core UltraSPARC T1 processor, supports
up to 64 GiB of ECC DDR2 SDRAM system memory using 16 DIMM slots, and up to 584 gigabytes
of internal storage with Serial attached SCSI drives.[3] InfoWorld noted that the system is light for
its class.[4] It is 3.5 inches tall, 16.7 inches wide, 24.4 inches deep and weighs 40 pounds.

Operating Systems
The T2000 was offered by Sun preloaded with the Solaris 10 operating system

Sun Fire X4500

The Sun Fire X4500 data server integrates server and storage technologies. It was announced in
July, 2006[1] and is part of the Sun Fire server line from Sun Microsystems.
In July 2008, Sun announced the X4540 model (code-named Thor), which doubles the processing
power of the X4500.
In November 2010, Oracle designated that the X4540 is end-of-life and has no next-generation
replacement model.

Hardware
The Sun Fire X4500 supports two dual-core AMD Opteron processors and up to 64 GB RAM. With
forty-eight 500/1000/2000 GB SATA drives, it provides up to 96 TB of raw storage in four rack
units.
The Sun Fire X4540 supports two quad- or six-core AMD K10 (Barcelona) processors and up to
128 GB RAM. The new model also uses PCI Express IO technology, and added a compact
flash disk slot for booting the operating system.
A significant feature of both systems is that the I/O framework was designed to handle high
throughput on all disks simultaneously. These were the first systems designed specifically
with ZFS in mind, so no hardware RAID is included.
Supported operating systems
 Solaris 10
 Red Hat Enterprise Linux 4.0
 SUSE Linux Enterprise Server 10
 Rocks Cluster Distribution[4]
 Microsoft Windows Server 2003 (Standard and Enterprise)
 Microsoft Windows Server 2008[6]

Sun Netra
The Sun Netra brand has been used for a variety of server computers from Sun
Microsystems since 1994. The original Netra servers (such as the Netra i and Netra j models) were
re-badged SPARCstation and Sun Ultra series systems bundled with (web) server application
software.

Sun Fire
Sun Fire is a series of server computers introduced in 2001 by Sun Microsystems (since 2010,
part of Oracle Corporation). The Sun Fire branding coincided with the introduction of
the UltraSPARC III processor, superseding the UltraSPARC II-based Sun Enterprise series. In
2003, Sun broadened the Sun Fire brand, introducing Sun Fire servers using the
Intel Xeon processor. In 2004, these early Intel Xeon models were superseded by models powered
by AMD Opteron processors. Also in 2004, Sun introduced Sun Fire servers powered by
the UltraSPARC IV dual-core processor. In 2007, Sun again introduced Intel Xeon Sun Fire
servers, while continuing to offer the AMD Opteron versions as well.
SPARC-based Sun Fire systems were produced until 2010, while x86-64 based machines were
marketed until mid-2012. In mid-2012, Oracle Corporation ceased to use the Sun Fire brand for
new server models.

Operating systems
UltraSPARC-based Sun Fire models are licensed to run the Solaris operating system versions 8,
9, and 10. Although not officially supported, some Linux versions are also available from third
parties, as well as OpenBSD and NetBSD.
Intel Xeon and AMD Opteron based Sun Fire servers support Solaris 9 and 10, OpenBSD, Red
Hat Enterprise Linux versions 3 - 6, SUSE Linux Enterprise Server 10 and 11, Windows
2000, Windows Server 2003, 2008, and 2008 R2 .

Model nomenclature
Later Sun Fire model numbers have prefixes indicating the type of system, thus:

 V: entry level and mid-range rackmount and cabinet servers (UltraSPARC, IA-32 or AMD64)
 E: high-end enterprise class cabinet servers with high-availability features (UltraSPARC)
 B: blade servers (UltraSPARC or IA-32)
 X: rackmount x86-64 based servers
 T: entry level and mid-range rackmount servers based on UltraSPARC T-series

When Sun offered Intel Xeon and AMD Opteron Sun Fire servers under the V-Series sub brand,
Sun used an x suffix to denote Intel Xeon processor based systems and a z suffix for AMD Opteron
processor based systems, but this convention was later dropped.
Sun's first-generation blade server platform, the Sun Fire B1600 chassis and associated blade
servers, was branded under the Sun Fire server brand. Later Sun blade systems were sold under
the Sun Blade brand. In 2007, Sun, Fujitsu and Fujitsu Siemens introduced the common SPARC
Enterprise brand for server products. The first SPARC Enterprise models were the Fujitsu-
developed successors to the midrange and high-end Sun Fire E-series. In addition, the Sun Fire
T1000 and T2000 servers were rebranded as the SPARC Enterprise T1000 and T2000 and sold
under the Fujitsu brands, although Sun continued to offer these with their original names. Later T-
series servers have also been badged SPARC Enterprise rather than Sun Fire.
Since late 2010, Oracle Corporation no longer uses Sun Fire brand for their current T series
SPARC servers, and since mid-2012 for new X series x86-64 machines based on Intel Xeon CPUs.
x86-64 server models which had been developed by Sun Microsystems before its acquisition, and
were still in production, have all been rebranded as Sun Server X-series.

Sun Fire model range

Some servers were produced in two versions, the original version and a later RoHS version. Since
a general maintenance and upgrade guideline is that RoHS components and spares may be
installed into the original non-RoHS versions of that server, the end-of-life (EOL) date of a server
is deemed the EOL date of the RoHS version of that server in this listing.

UltraSPARC architecture
x86/x64 architecture
Sun Server / Oracle Server
As of 2012, the x86 server range continued under the "Sun Server" or "Oracle Server" names.
 Sun Blade
Sun Blade is a line of blade server computer systems sold by Sun Microsystems from 2006
onwards.
In June 2006, Sun announced the AMD Opteron-based Sun Blade 8000 modular blade server
system. The Sun Blade 8000 chassis can hold up to 10 Sun Blade X8420 or X8440 modules.
The following year, the Sun Blade 6000 system was launched. This allowed mixed UltraSPARC
and x64 (x86-64) architecture blades. The Sun Blade T6300 and T6320 modules run Solaris and
use UltraSPARC T1 and UltraSPARC T2 processors respectively, while the Sun Blade X6220 and
X6250 modules use AMD Opteron or Intel Xeon 5000-series processors respectively.
The X-series blades support Solaris, Oracle Linux, RHEL, SLES, Windows Server or VMware.

Cobalt Qube
The Cobalt Qube is a computer server appliance product line developed by Cobalt Networks, Inc.
(later purchased by Sun Microsystems) from 1998 to 2002 featuring a modified Red Hat
Linux operating system and a proprietary GUI for server management. The original Qube systems
were equipped with RM5230 or RM5231 microprocessors but later models used AMD K6-
2 chips. NetBSD operating system has been ported to both the Cobalt Qube and RaQ

Models
The Qube 2700 was the first product released by Cobalt Networks in 1998.[1] Mark Orr, one of the
Cobalt Networks' CEOs, came up with the cobalt color. The green LED in the front was Bill Scott's
idea. The 2700 was not a development version number but came from the atomic number of cobalt,
27. The Qube 2700 used the RM5230 microprocessor.
The next product was called the Qube 2800 before being sold. But, released in 2000, was
eventually called the Qube 2, leaving the 2800 to designate the system type. The Qube 2 used
the RM5231 microprocessor.
Under an OEM arrangement, the Qube 2 units were also produced by Gateway in the form of the
Gateway Micro Server. The casing featured on these units was black instead of cobalt blue.
The Qube 3, released in 2002, used an AMD K6-2 CPU at either 300 MHz or 450 MHz and was
the last product in the Qube line.
A fourth Qube model was in development but was never released. However, several models were
released in the data center-friendly Cobalt RaQ product line after the Cobalt Qube was
discontinued.

Cobalt RaQ
The Cobalt RaQ is a 1U rackmount server product line developed by Cobalt Networks, Inc. (later
purchased by Sun Microsystems) featuring a modified Red Hat Linux operating system and a
proprietary GUI for server management. The original RaQ systems were equipped
with MIPS RM5230 or RM5231 CPUs but later models used AMD K6-2 chips and then eventually
Intel Pentium III CPUs for the final models.
The Cobalt RaQ was the second product line produced by Cobalt Networks; the first was the Cobalt
Qube.

Server specifications
Below is a list of Cobalt RaQ types, and their specs.

There were variants of the RaQ 3 and RaQ 4 models known as the RaQ 3i or RaQ 4i (SCSI support,
two Ethernet connectors, PCI connector), and the RaQ 4r (SCSI support, two Ethernet connectors,
and RAID). RAID on these models was accomplished in software using a second IDE channel on
the motherboard for the second hard drive. There was also a "bare bones" RaQ 4 model that had
a single Ethernet adapter, no external SCSI, and a single hard drive.

Cisco Server
Cisco Unified Computing System
The Cisco Unified Computing System (UCS) is a data center server computer product line
composed of computing hardware, virtualization support, switching fabric, and management
software introduced in 2009 by Cisco Systems.[1][2] The products are marketed for scalability by
integrating many components of a data center that can be managed as a single unit.

Computing
The UCS product line was announced in March and first installed in September 2009, at the
customer premises of Tutor Perini. The first customer purchase of UCS was by Fiserv, a financial
services customer in Milwaukee, Wisconsin. The entire project was code named "California".The
computing component of the UCS is available in two versions: the B-series (a powered chassis
and full and/or half slot blade servers), and the C-series for 19-inch racks (that can be used with
fabric interconnects). The computer hardware managed by the UCS Manager software on the
fabric Interconnects can be any combination of the two. Both form factors use standard
components, including Intel x86-64 processors and DIMM memory. The servers are marketed
with converged network adapters and port virtualization. Around 2010, an extended memory
technology expanded the number of memory sockets that can be connected to a single memory
channel in some models. A fifth generation was announced in July 2018.

Virtualization
Cisco UCS supports several hypervisors including VMware ESXi, Microsoft Hyper-V and Citrix
Systems' Xen server.[7] VMware provides a special version of ESXi.
In 2017 Cisco announced a partnership with Docker to include its enterprise software in UCS
products to directly provide operating-system-level virtualization (containerization)

Networking
The Cisco fabric interconnects (FI) provide network connectivity for the chassis, blade servers and
rack servers connected to it through different speeds of Ethernet and Fiber Channel over
Ethernet (FCoE). The fabric interconnects are derived from the Nexus 5500 and 9300 series switch
families and run NX-OS as well as UCS Manager software. FCoE allows connection to storage
area network storage, To allow stateless compute in addition to local storage capacity. Cisco has
produced the following series for fabric interconnects: 6100 Series (discontinued), 6200 Series,
6300 Series, and 6400 Series. The fabric interconnect can further connect to multiple fabric
extenders (FEX), into what they call a "unified fabric".
By 2012, analysts noticed that Fibre Channel vendors such as Brocade Communications
Systems had already lost market share to Cisco

Management
Management of the system devices is handled by the UCS Manager software integrated with fabric
interconnects. Virtual machines can be moved from one physical chassis to another, applications
may be moved between virtual machines, and management may even be conducted remotely from
an iPhone using SiMU - Simple iPhone Management of UCS. In addition to the embedded
software, administrators may also manage the system from VMware's vSphere. The Cisco
Integrated Management Controller (CIMC) is used to configure and manage C-Series servers not
connected to a manager. The CIMC can also be used to manage B-series blades in addition to the
UCS Manager application if configured.
In November 2012, Cisco announced UCS Central which extends management across multiple
domains of UCS,and then something called UCS Director. A security flaw in UCS Central was
disclosed when it was patched in 2016.
Many configuration details are set up on the UCS manager in a service profile and applied to the
servers. Cisco offers a UCSM Platform Emulator, where the full logical configuration of a server
can be created from the user interface or the API methods, and later applied to the physical
hardware.
In 2017, Cisco announced a cloud based management offering for UCS infrastructure called Cisco
Intersight. This platform augmented the UCS management platform.

Servers
 Cisco UCS B-Series Blade Servers (Mini, B460 M4, B420 M4, B260 M4, B200 M4, B480 M5,
B200 M5)
 Cisco UCS C-Series Rack Servers (C460 M4, C240 M4, C220 M4, C480 M5, C240 M5, C220
M5),
 Cisco UCS Fabric Interconnects (6200 series, 6300 series),