HCIA-Cloud Computing V5.
0 Learning Guide Page 29
2 Server Basics
Servers are the foundation of all service platforms, including cloud computing platforms.
But what is a server? What are the key technologies for servers? Let's find the answers in
this chapter, and start our learning journey into cloud computing.
2.1 Introduction to Servers
2.1.1 What Is a Server?
[Link] Server Definition and Features
A server is a type of computer. It runs faster, carries more loads, and costs more than
ordinary computers. A server provides services to users. There are file servers, database
servers, and application servers.
A server is a mainstream computing product developed in 1990s. It can provide network
users with centralized computing, information release, and data management services. In
addition, a server can share drives, printers, and modems to which the server is
connected, and dedicated communication devices with network users. So far, general-
purpose servers are still the most widely used basic IT devices in enterprises, accounting
for 90% or even higher of enterprise IT computing devices.
Servers are similar to PCs, but with features PCs do not have.
A server has the following features:
⚫ R: Reliability – the duration that the server operates consecutively
⚫ A: Availability – percentage of normal system uptime and use time
⚫ S: Scalability – including hardware expansion and operating system (OS) support
capabilities
⚫ U: Usability – easy to maintain and restore server hardware and software
⚫ M: Manageability – monitoring and alarm reporting of server running status, and
automatically intelligent fault processing
The preceding features are necessary for a qualified server.
HCIA-Cloud Computing V5.0 Learning Guide Page 30
[Link] Server Application Scenarios
Figure 2-1 Server use cases
Servers have been widely used in various fields, such as the telecom, government,
finance, education, enterprise, and e-commerce. Servers can provide users with the file,
database, email, and web services.
As shown in Figure 2-1, there are two types of server application deployment
architectures:
⚫ C/S: short for Client/Server. In this architecture, the server program runs on the
server, and the client software is installed on the client. The server and client perform
different tasks. The client carries the front-end GUI and interaction operations of
users, and the server processes the background service logic and request data. This
greatly improves the communication speed and efficiency between the two ends. For
example, you can install the vsftpd program on a file server and start the service.
After you install the FileZilla or WinSCP client on your PC, you can upload and
download files using the client.
⚫ B/S: short for Browser/Server. In this architecture, users only need to install a browser.
The application logic is centralized on the server and middleware, which improves
the data processing performance. For example, when accessing a website, we only
need to enter the domain name of the website in the browser, for example,
[Link]. Then we can see the web services provided by the background
servers of the website. We do not need to care the background servers that provide
services, such as the database service, proxy service, and cache service.
HCIA-Cloud Computing V5.0 Learning Guide Page 31
2.1.2 Server Development History
Figure 2-2 Server development history
Figure 2-2 shows the four phases of server development.
⚫ Mainframe and midrange computers
In the 1940s and 1950s, the first generation of vacuum tube computers emerged. The
computer technology develops rapidly from vacuum tube computers, transistor
computers, integrated circuit computers, to large-scale integrated circuit computers.
In the 1960s and 1970s, mainframes were scaled down for the first time to meet the
information processing requirements of small- and medium-sized enterprises and
institutions. The cost was acceptable.
⚫ Microcomputers
In the 1970s and 1980s, mainframes were scaled down for the second time. Apple Inc.
was founded in 1976, and launched Apple II in 1977. In 1981, IBM launched IBM-PC.
After several generations of evolution, it occupied the personal computer market and
made personal computers popular.
⚫ x86 servers
In 1978, Intel launched the first-generation x86 architecture processor, 8086
microprocessor.
In 1993, Intel officially launched the Pentium series, which brought the x86 architecture
processor to a new level of performance.
In 1995, Intel launched Pentium Pro, the x86 processor for servers, ushering in the x86
era. The standardization and openness of Pentium Pro also promoted the market
development and laid a solid foundation for the cloud computing era.
⚫ Cloud computing
Since 2008, the concept of cloud computing has gradually become popular. Cloud
computing is regarded as a revolutionary computing model because it enables the free
flow of supercomputing capabilities through the Internet. Enterprises and individual users
do not need to purchase expensive hardware. Instead, they can rent computing power
through the Internet and pay only for the functions they need. Cloud computing allows
users to obtain applications without the complexity of technologies and deployment.
Cloud computing covers development, architecture, load balancing, and business models,
and is the future model of the software industry.
The computing industry has developed for nearly half a century and continuously
changed other industries. The computing industry itself is evolving.
In the early mainframe and midrange computer era, dedicated computing is used, which
is called computing 1.0. In the x86 era, under the leadership of Intel and driven by
Moore's Law, computing has shifted from dedicated to general-purpose. A large number
of data centers have emerged, which is called computing 2.0. With the rapid
development of digitalization, the world is developing towards intelligent. Computing is
HCIA-Cloud Computing V5.0 Learning Guide Page 32
not limited to data centers, but also enters the full-stack all-scenario (computing 3.0) era.
This era is featured by intelligence, so it is also called intelligent computing.
2.1.3 Server Types
[Link] Server Classification - Hardware Form
Figure 2-3 Server classification - Hardware form
As shown in Figure 2-3, servers can be classified into the following types based on
hardware forms:
⚫ Tower server
Some tower servers use a chassis roughly the same size as an ordinary vertical computer,
while others use a large-capacity chassis, like a large cabinet.
⚫ Rack server
A rack server looks different from a computer, but looks similar to a switch. The
specifications of a rack server include 1 U (1 U = 1.75 inches), 2 U, and 4 U. A rack server
is installed in a standard 19-inch cabinet. Most of the servers in this structure are
functional servers. A rack server is usually small in size. Multiple servers can be placed in
a cabinet at the same time to obtain a higher processing capability.
⚫ Blade server
Each blade server is a plugboard equipped with processors, memory modules, hard drives,
and related components. Due to the special architecture, blade servers require dedicated
chassis. Generally, a chassis can hold several to dozens of blade servers, suitable for
scenarios such as high-performance computing, front-end servers running multiple
applications, and backend central databases.
For details about mainframes and midrange computers, see the preceding description.
HCIA-Cloud Computing V5.0 Learning Guide Page 33
[Link] Server Classification - Service Scale
Figure 2-4 Server classification - Service scale
As shown in Figure 2-4, servers can be classified into entry-level servers, work group
servers, department-level servers, and enterprise-level servers based on the service scale.
2.1.4 Server Hardware
[Link] Hardware Structure
Figure 2-5 Hardware structure
Figure 2-5 shows the hardware structure of a Huawei TaiShan 200 server. You can learn
about the internal components of the server and their locations. Important hardware in
the server, such as the CPU, memory, and drive, will be described in detail later.
[Link] CPU
[Link].1 CPU Definition and Components
The Central Processing Unit (CPU) is the computing and control core of a computer.
The CPU is the core processing unit on a server, and a server is an important device on
the network and needs to process a large number of access requests. Therefore, servers
must have high throughput and robust stability, and support long-term running. The CPU
is the brain of a computer and is the primary indicator for measuring server performance.
The CPU, internal storage, and input/output devices are key components of a computer.
The CPU interprets computer instructions and processes computer software data.
HCIA-Cloud Computing V5.0 Learning Guide Page 34
Figure 2-6 CPU structure
As shown in Figure 2-6, the middle part is the CPU structure. The CPU consists of a logic
operation unit, a control unit, and a storage unit.
⚫ The computer controls the entire computer according to a pre-stored program, and
the program refers to an instruction sequence that can implement a function. The
controller is an organization that issues commands to various logic circuits according
to the instructions. The controller is a command center of the computer, controls
work of an entire CPU, and determines automation of a running process of the
computer.
⚫ The ALU is a part of a computer that performs a variety of arithmetic and logical
operations. Basic operations of an ALU include arithmetic operations such as
addition, subtraction, multiplication, and division, logical operations such as AND,
OR, NOT, and XOR, and other operations such as shift, comparison, and transfer. The
ALU is also called the arithmetic logic component.
⚫ The register is used to temporarily store the data involved in operations and the
operation results. It can receive, store, and output data.
[Link].2 CPU Frequency
Generally, the following CPU frequency parameters are used to measure CPU
performance:
⚫ The dominant frequency is also called clock speed. It indicates, in MHz or GHz, the
frequency at which a CPU computes and processes data.
⚫ The external frequency is the reference frequency of a CPU, measured in MHz. The
CPU external frequency determines the speed of the motherboard.
⚫ The bus frequency directly affects the speed of data exchange between a CPU and a
dual in-line memory module (DIMM).
⚫ The multiplication factor is the ratio of the dominant frequency to the external
frequency.
HCIA-Cloud Computing V5.0 Learning Guide Page 35
[Link] Memory
Figure 2-7 Memory
The storage, an important computer component, is used to store programs and data. For
computers, the memory function can be supported and normal working can be ensured
only when the storage is available. Storage is classified, by purpose, into main memory
and external storage. Main memory, referred to as internal storage, is the storage space
that the CPU can address.
As a main computer component, the memory is in opposition to the external storage.
Programs, such as the Windows OS, typing software, and game software, are usually
installed on external storage devices such as drives. To use these programs, you must
load them into the memory. Actually, the memory is used when we input a piece of text
or play a game. Bookshelves and bookcases for putting books are just like the external
storage, while the desk is like the memory. Generally, we store large volumes of data
permanently in the external storage and store small volumes of data and a few programs
temporarily in the memory.
The memory, one of important computer components, communicates with the CPU. The
memory consists of the memory chip, circuit card, and edge connector.
Comply with the following principles when installing DIMMs on servers:
⚫ DIMMs on the same server must be of the same model.
⚫ At least one DIMM must be configured in slots supported by CPU 1.
⚫ Optimal memory performance can be achieved if the processors in a server are
configured with the same number of DIMMs and the DIMMs are evenly distributed
among the memory channels. Unbalanced configuration impacts memory
performance and is not recommended.
[Link] Drive
The drive is the most important storage device of a computer.
The drive interface, connecting a drive to a host, is used to transmit data between the
drive cache and the host memory. The drive interface type determines the connection
HCIA-Cloud Computing V5.0 Learning Guide Page 36
speed between the drive and the computer, how quickly programs run, and overall
system performance.
Figure 2-8 Drive types
Common drive types include SATA, SAS, NL-SAS, and SSD. The preceding table lists their
characteristics. MTBF is short for Mean Time Between Failures. A larger value indicates a
lower failure rate of the drive. SATA and NL-SAS drives are cheaper, SAS drives are more
expensive, and SSDs are the most expensive.
[Link] RAID Controller Card
[Link].1 RAID Overview
Generally, a server has a RAID controller card. The RAID controller card is also called the
disk array card. To understand a RAID controller card, you need to know what RAID is.
Redundant Array of Independent Disks (RAID) is a data storage virtualization technology
that combines multiple physical disk drive components into one or more logical units for
the purposes of data redundancy, performance improvement, or both. For details about
the working principles of RAID, see the chapter of storage basics.
The RAID controller card provides the following functions:
⚫ Combines several drives into a system that is managed by the array controller
according to certain requirements.
⚫ Improves the performance and reliability of the disk subsystem.
HCIA-Cloud Computing V5.0 Learning Guide Page 37
Figure 2-9 LSI SAS3108 RAID controller card
[Link].2 RAID Hot Spare and Reconstruction
Hot spare definition: If a drive in a RAID array fails, a hot spare is used to automatically
replace the failed drive to maintain the RAID array's redundancy and data continuity.
Hot spare is divided into the following two types:
⚫ Global: The spare drive is shared by all RAID arrays in the system.
⚫ Dedicated: The spare drive is used only by a specific RAID array.
A0 Reconstruction XOR
Failure
Reconstruction A0
Replace A1 A2 P
ment
Data drive Data drive Data drive Parity drive
Hot spare
drive
Figure 2-10 Hot spare and reconstruction
The process of restoring data from a faulty data drive to a hot spare drive is called data
reconstruction. Generally, the data parity mechanism in RAID is used to reconstruct data.
Data parity: Redundant data is used to detect and rectify data errors. The redundant data
is usually calculated through Hamming check or XOR operations. Data parity can greatly
improve the reliability, performance, and error tolerance of the drive arrays. However, the
system needs to read data from multiple locations, calculate, and compare data during
the parity process, which affects system performance.
Generally, RAID cannot be used as an alternative to data backup. It cannot prevent data
loss caused by non-drive faults, such as viruses, man-made damages, and accidental
deletion. Data loss here refers to the loss of OS, file system, volume manager, or
application system data, not the RAID data loss. Therefore, data protection measures,
HCIA-Cloud Computing V5.0 Learning Guide Page 38
such as data backup and disaster recovery, are necessary. They are complementary to
RAID, and can ensure data security and prevent data loss at different layers.
[Link].3 RAID Implementation - Hardware
Figure 2-11 RAID implementation - Hardware
As shown in Figure 2-11, hardware RAID is implemented using a hardware RAID adapter
card.
The hardware RAID can be a built-in or external RAID.
A RAID controller card has a processor inside and can control the RAID storage
subsystem independently from the host. The RAID controller card has its own
independent processor and memory. It can calculate parity information and locate files,
reducing the CPU computing time and improving the parallel data transmission speed.
[Link].4 RAID Implementation - Software
Software RAID implements RAID functions by installing software on the OS.
Software RAID has the following characteristics:
⚫ Software RAID does not require expensive RAID controller cards, reducing the cost.
⚫ RAID functions are performed by CPUs, requiring significant CPU resources, such as
for large numbers of RAID 5 XOR operations.
Compared with hardware RAID, software RAID has other defects. For example, software
RAID does not support the following functions:
⚫ Hot swap of drives
⚫ Drive hot spare
⚫ Remote array management
⚫ Support for bootable arrays
⚫ Array configuration on drives
⚫ S.M.A.R.T. for drives
[Link].5 RAID Implementation - Mode Comparison
Figure 2-12 compares software RAID and hardware RAID.
HCIA-Cloud Computing V5.0 Learning Guide Page 39
Figure 2-12 RAID implementation - Mode comparison
[Link] NIC
A network interface card (NIC or network adapter) is an indispensable part of a
computer network system. An NIC enables a computer to access networks.
For a server, the NIC provides the following functions:
⚫ Fixed network address
⚫ Data sending and receiving
⚫ Data encapsulation and decapsulation
⚫ Link management
⚫ Encoding and decoding
Huawei servers have the following four types of NICs:
⚫ LOM card
It is embedded directly into the PCH chip on the server motherboard and cannot be
replaced.
It provides two external GE electrical ports + two 10 Gbit/s optical/electrical ports. LOM
cards do not occupy PCIe slots.
⚫ PCIe card
Huawei has both self-developed and purchased PCIe cards. They can be installed in
standard PCIe slots.
⚫ FlexIO card
Huawei-developed, non-standard PCIe card, which can only be used with Huawei rack
servers.
⚫ Mezzanine card
Mezzanine cards are only used on the compute nodes of Huawei E9000 blade servers.
HCIA-Cloud Computing V5.0 Learning Guide Page 40
[Link] PSU and Fan Module
Figure 2-13 PSU and fan module
As shown in Figure 2-13, the power supply unit (PSU) and fan module support the power
load and heat dissipation of the server. Redundant PSUs and fan modules are deployed
on servers to ensure continuous running of the system.
The following power supply redundancy modes are available:
⚫ 1+1: In this mode, each module provides 50% of the output power. When one
module is removed, the other provides 100% of the output power.
⚫ 2+1: In this mode, each module provides 1/3 of the output power. When one module
is removed, each of the other two modules provides 50% of the output power.
2.1.5 Key Server Technologies
[Link] BMC
[Link].1 What Is IPMI?
The Intelligent Platform Management Interface (IPMI) is a set of open and standard
hardware management interface specifications that defines specific methods for
communication between embedded management subsystems.
The IPMI is an industrial specification used for peripherals in Intel-based enterprise
systems. This interface specification was laid down by Intel, HP, NEC, Dell, and
SuperMicro. Users can use the IPMI to monitor the physical health status of servers, such
as the temperature, voltage, fan status, and power status. Moreover, the IPMI is a free
specification. Users do not need to pay for this specification.
IPMI development:
⚫ In 1998, Intel, DELL, HP, and NEC put forward the IPMI specification. The
temperature and voltage can be remotely controlled through the network.
⚫ In 2001, the IPMI was upgraded from version 1.0 to version 1.5. The PCI
Management Bus function was added.
⚫ In 2004, Intel released the IPMI 2.0 specification, which is compatible with the IPMI
1.0 and 1.5 specifications. Console Redirection is added. Servers can be remotely
HCIA-Cloud Computing V5.0 Learning Guide Page 41
managed through ports, modems, and LANs. In addition, security, VLANs, and blade
servers are supported.
IPMI information is exchanged using the baseboard management controller (BMC).
Entry-level intelligent hardware, not the OS, handles management.
Figure 2-14 IPMI
The IPMI management tool on the client remotely manages the server through the IPMI
port on the server.
[Link].2 BMC
The BMC complies with the IPMI specification. It collects, processes, and stores sensor
signals, and monitors component operating status. It supplies the chassis management
module with managed objects' hardware status and alarm information. The management
module uses this information to manage the devices.
The BMC provides the following functions:
⚫ Remote control
⚫ Alarm management
⚫ Status check
⚫ Device information management
⚫ Heat dissipation control
⚫ Support for IPMItool
⚫ Web-based management
⚫ Centralized account management
[Link].3 iBMC
The Huawei Intelligent Baseboard Management Controller (iBMC) is a Huawei
proprietary embedded server management system designed for the whole server lifecycle.
HCIA-Cloud Computing V5.0 Learning Guide Page 42
Figure 2-15 iBMC
The iBMC provides a series of management tools for hardware status monitoring,
deployment, energy saving, and security, and standard interfaces to build a
comprehensive server management ecosystem. The iBMC uses Huawei-developed
management chip Hi1710 and multiple innovative technologies to implement refined
server management.
The iBMC provides a variety of user interfaces, such as the CLI, web-based user interface,
IPMI integration interface, SNMP integration interface, and Redfish integration interface.
All user interfaces adopt the authentication mechanism and high-security encryption
algorithm to enhance access and transmission security.
[Link] BIOS
The Basic Input/Output System (BIOS) is a system's foundation: a group of programs
providing the most direct control of system hardware.
The BIOS is a bridge between the system kernel and the hardware layer. The BIOS is a
small system at the motherboard level. It initializes the hardware of the system (mainly
the motherboard), such as the CPU, memory, drive, keyboard, graphics card, and NIC.
Initialization is the main task of the BIOS. The BIOS on a traditional PC has an int19
software interrupt function. After the initialization is complete, the BIOS enters the int19
interrupt mode to search for the boot medium, such as a floppy disk, CD-ROM, hard
drive, flash memory, or network, reads the content of the first sector to 0000:7C00 in the
memory, and goes to this address. int19 is an OS bootloader. Therefore, the BIOS can
boot the OS. Of course, many BIOS functions have been added, for example, ACPI
interface for power management, USB driver, PXE network boot function, drive
encryption, TPM interface, BIOS configuration interface, and BIOS automatic recovery.
The BIOS provides the following functions:
⚫ Hardware detection and initialization
⚫ OS boot
⚫ Advanced power management
HCIA-Cloud Computing V5.0 Learning Guide Page 43
2.2 Quiz
Which types of hardware are critical in the hardware structure of the server? What are
their functions?