UCU Data Management and Analysis

M1: Understanding Data

 Foreword


Data carries information during the transmission on networks. What
is the relationship between information and data? What is the
function of data management? This course describes the definition
of information and data in the computer field, their relationship, as
well as the concept, development history, and development trend of
data storage.

2 Huawei Confidential
 Objectives

On completion of this course, you will be able to understand:


Definition of information and data

Concept of data storage

History of data storage

Development trend of data storage products

3 Huawei Confidential
 Contents

1. Data and Information

2. Data Storage
3. Development of Storage Technologies
4. Development Trend of Storage Products

4 Huawei Confidential
What Is Data

SNIA (Storage Networking Industry Association) defines data as the
digital representation of anything in any form.
Format in which data is stored
0101001010100010000
Email 0111100011100010001
1100011100000111101
0101001010101001010
Digital music 1001010010101001010
1010010101000101010
0101010010101010100
Digital video
1010101010100101000
1001010010101010101
Eboo 0010101010010101010
k 1010010101010100101
0010100100010101010
01001010010010
5 Huawei Confidential
Data Types

Semi-
Structured structured Unstructur
data data ed data

6 Huawei Confidential
Types of Data

State Data

Historical Data

Solu-
tions

State data records the current state of data.
Storage soft-

Eg Check your
ware Bank Account: BALANCES withdraws, Online
games where progress is stored as per what is happening
 Storagerecorded
Basically hard- on Paper before computers came
ware

It is now readily stored on some computers

Also use computers to automate processes
7 Huawei Confidential
Types of Data

State Data

Historical Data

Solu-
tions
Historical data records the events that already happened.
Storage soft-
-
What products
ware did a user purchase?
-
Cell Phone records
Storage hard-
-
We do notware
change this type of data.
-
Current Systems generate large amounts of historical data

8 Huawei Confidential
Data Processing Cycle

Data processing is the reorganization or reordering of data by humans or
machines to increase their specific value. A data processing cycle
includes three basic steps: input, processing, and output.

Processin
Input Output
g

9 Huawei Confidential
What Is Information

Information is processed, structured,
or rendered in a given context to Infor
Processing matio
make it meaningful and useful. n


Information is processed data,
D
including data with context, at
relevance, and purpose.
involves the manipulation of raw data.
It also
a Data
Redun
dancy

10 Huawei Confidential
Data vs. Information

After being processed, data can be converted into information.

In order to be stored and transmitted in IT systems, information needs to
be processed as data.

Item Data Information

Raw and meaningless, with no
Feature Valuable and logical
specific purpose
Essence Original materials Processed data
Data never depends on
Dependence Information depends on data
information
Meteorological data or satellite
Example Weather forecasts
image data

11 Huawei Confidential
Information Lifecycle Management

Information lifecycle management (ILM) refers to a set of management
theories and methods from the stage when the information is generated
and initially stored to the stage when the information is deleted.

Data
value

Creation Protection Access Migration Archiving Destruction

12 Huawei Confidential
 Contents

1. Data and Information

2. Data Production Methods
3. Development of Storage Technologies
4. Development Trend of Storage Products

13 Huawei Confidential
What Is Data Production Methods

14 Huawei Confidential
Traditional Production methods

Understand the needs of users based on requirements

Make a product based on that

Sometimes mistakes are bound to happen
Solu-
tions

Process can be significantly slow
Storage soft-
ware

Storage hard-
ware

15 Huawei Confidential
Data Driven Production methods

Give products to users

Aim to collect user data for example how are they using your application

Has iterations because you are in a constant cycle of feedback
Solu-
tions

Improves the efficiency of our production
Storage soft-
ware

Storage hard-
ware

16 Huawei Confidential
Data Driven Production methods cont

Take an example of a supermarket; How products are placed in a
supermarket?

Take note of the Product placement on shelves.
Solu-

Purchase tions
habits of customers

Storage soft-
ware

Storage hard-
ware

17 Huawei Confidential
Data Driven Production methods cont

Take an example of a Education industry

Using applications or websites.

How are the students interacting with the data.
Solu-
tions

Data collected from platforms improves the learning experience
Storage soft-
ware

Storage hard-
ware

18 Huawei Confidential
Physical Structure of Storage
Hosts

Network switching devices NAS gateway

Storage application
software
Front-end board of
controllers

Cache
Storage
Storage
system Back-end board of management
controllers software

Connectors

Disks

19 Huawei Confidential
Data Storage Types

Internal DAS
Storage SAN
Storage FAS
External NAS
Storage

20 Huawei Confidential
Evolution of Data Management Technologies
Distributed
Dedicated storage device
storage
device
Tape and
single disk
Big data
management
Traditional
Card and database
paper tape File system system
management management

Manual
manageme
nt

21 Huawei Confidential
Data Storage Application

Database servers
Users
Application
servers

Email servers

Offline
storage
File servers Online Nearline storage
Employee storage
s Classification based on data access frequencies

Data Data Data

generation processing management

22 Huawei Confidential
 Contents

1. Data and Information

2. Data Storage
3. Development of Storage Technologies
 Storage Architecture
▫ Storage Media

▫ Interface Protocols

4. Development Trend of Storage Products

23 Huawei Confidential
History of Storage Architecture Development

2000s
1990s • Distributed
• Storage storage
1980s network • Cloud
• External storage
storage
1950s
• Traditional
storage

24 Huawei Confidential
From Disks to Disk Arrays

Disks in a server Early external storage Storage arrays

Server Server Server

CPU CPU CPU

RAM RAM RAM

Disks Disks Disks

JBOD
Controllers

25 Huawei Confidential
From Separation to Convergence

SAN NAS Clients Unified

Storage
Server Server
File Ethernet
syste Server
m

SAN LAN

File NAS SAN

syste
m Storage
Storage File
syste
m Storage

26 Huawei Confidential
Distributed Storage

Physical resources are organized
using software to form a high-
Front-end
performance logical storage pool, service
network Management
ensuring reliability and providing network

multiple storage services.


Generally, distributed storage
scatters data to multiple
independent storage servers in a
...
scalable system structure. It uses
those storage servers to share Node Node Node Node

storage loads and uses location Back-end storage

network
servers to locate storage
27 Huawei Confidential
information.
Storage Virtualization

Storage virtualization consolidates Clients or servers

the storage devices into logical

resources, thereby providing
comprehensive and unified Management
storage services.
Logical resources

Unified functions are provided
regardless of different storage
Virtualization
forms and device types.
Physical resources

28 Huawei Confidential
Cloud Storage

The cloud storage system comprises storage devices, applications, and services. It uses
highly virtualized multi-tenant infrastructure to provide scalable storage resources,
which can be dynamically configured based on organization requirements.

Legacy New
applications applications

Protocol and interface

Converged
Elastic Block storage service File storage service Object storage service

Open Storage device and storage node

Universal hardware platforms

29 Huawei Confidential
 Contents

1. Data and Information

2. Data Storage
3. Development of Storage Technologies
▫ Storage Architecture
 Storage Media
▫ Interface Protocols

4. Development Trend of Storage Products

30 Huawei Confidential
History of HDDs

Larger capacity with the smaller size.

1950s– 1960s– 1970s– 1980s–

1990s–
1960s 1970s 1980s 1990s
Microdrive
Giant disk 14-inch disk 8-inch disk Portable
IBM 350 IBM 3340 Disks started 5.25-inch 1.8-inch
RAMAC to enter the PC 3.5-inch Hitachi: 1-inch
IBM 1301 market 2.5-inch Toshiba: 0.85-
inch
...

31 Huawei Confidential
History of SSDs

Solid-state drives (SSDs) were invented almost as early as HDDs, but
were unpopular due to high price in comparison to HDDs.

SSDs have gradually become more commercially accessible. With the
requirement for high access speed, the SSD market is booming.

• Bell Labs 1976 • SanDisk 1999 • Samsung 2010…

• Floating gate • Dataram sold launched • BiTMICRO laptops ran • Lower costs
transistor Bulk Core. flash SSDs. launched on SSDs. • Greater
• 2 MB • 20 MB flash SSDs. • 32 GB
capacity
capacity capacity • 18 GB capacity
1967 1991 capacity 2006

32 Huawei Confidential
Development of Flash Memory
Quad-level
Single-level cell (QLC):
cell (SLC): Triple-level larger and
fast cell (TLC): cheaper,
read/write mature applicable
and long mainstream to warm
service life technology data

Multi-level 3D TLC:
cell (MLC): improve the
moderate storage
read/write density
speed and through
service life multi-layer
overlaying
33 Huawei Confidential
 Contents

1. Data and Information

2. Data Storage
3. Development of Storage Technologies
▫ Storage Architecture

▫ Storage Media
 Interface Protocols

4. Development Trend of Storage Products

34 Huawei Confidential
Interface Protocols

Disk interfaces connect disks to hosts.

Interface protocols refer to the communication modes and requirements
that interfaces for exchanging information must comply with.

Parallel interface Serial interface

35 Huawei Confidential
History of Interface Protocols

100G NVMe-oF 400G

Ethernet NVMe-oF (FC-NVMe) Ethernet
Host 1.0 NVMe-oF 1.1
interface 2G FC (RDMA) (TCP)
iSCSI 4G FC 8G FC 16G FC 32G FC 64G FC

The first-gen FC
protocol was 2004 2008 2012 2016 2017 2018 2020…
released in
1994.

SATA
Disk interface SATA
1.0 4G FC 3.0
SAS 1.0 SAS 2.0 SAS 3.0 SAS 4.0
SATA 2.0

NVMe 1.0 NVMe 1.2 NVMe 1.3 NVMe 1.4

36 Huawei Confidential
NVMe and NVMe-oF

NVMe (Non-Volatile Memory
Express) CPU CPU

High performance
PCIe PCIe

Low latency
SAS controller

NVMe-oF (NVMe over Fabrics)
SAS expander

Potential: low latency and high
bandwidth

Purpose: accelerates data NVMe back-end SAS back-end
physical path physical path
transmission over the storage
network

37 Huawei Confidential
 Contents

1. Data and Information

2. Data Storage
3. Development of Storage Technologies
4. Development Trend of Storage Products

38 Huawei Confidential
History of Storage Products

Cloud
Media
HDD SSD

Architectur Tren
e d
Centralized Distributed

All-flash Intelligence
Manageme
nt
Manual O&M Intelligent O&M

39 Huawei Confidential
The Intelligence Era is Coming

Age of Steam Age of Age of Age of

Electricity Information Intelligence

Industry Cloud Service extension

computing AI
Data + Network coordination
Agricultur
e + Intelligence
Internet
IoT
Intelligent production

Service Big Customization

data

40 Huawei Confidential
Challenges to Data Storage
24/7 µs-level
Time-consuming Always-on Real-time, high-speed
RAID reconstruction services for all processing of
Service things production transactions
interruption due connected in the 5G era
to multiple
points of failure Finance High disk Reliabl
Fast
failure rate e
Gov & Large
Public enterpris
Low data e
deduplication
and Efficien
compression Large
ratio Carrier Transportation t

EB-level TCO reduction

Energy & Unstable latency Data monetized as Edge-center-cloud
Electricity under high assets and synergy,
Inefficient convergence of concurrency intelligent storage minimizing the per-
block, file, and object tiering, turning cold bit data cost and
services data to warm data maximizing the
to mine data value value

41 Huawei Confidential
Characteristics of Storage in the Intelligence Era
Intelligence
storage
Intelligent data
lifecycle management
Storage for AI AI in storage

Data convergence
Zero data migration Data collaboration
Convergence of Simplified protocol stack and
SAN and NAS data-storage convergence
Intelligent storage tiering of Compute-storage collaboration,
hot, warm, and cold data on-demand capacity expansion
Multi-service convergence Dedicated hardware and
(block, file, and object) distributed storage
Hardware + Algorithm + Architecture = Efficient storage
Robust reliability Large capacity High performance
Device-level of 99.99999% PB-level to EB-level ms-level to µs-level
System-level of 99.999% Efficient deduplication Intelligent distributed
& compression cache
Solution-level reliability + Cost-effective media, warm High-speed protocols
Cloud-level DR and cold data storage and algorithms
Minimum cost and maximum value for per-bit data

42 Huawei Confidential
Data Storage Trend
Distribute
d storage

Database Cloud
applications Blockchai
computing n AI
Intelligen
Mainframe
system t storage
Hyper- Big data
Private software
Traditional converged
and hardware
solutions with external infrastructure 5G
fully coupled storage (HCI) General-
computing, purpose
Brain-like
storage, and computing DNA Graph
Tensor computing
network computing computin computin
resources VM Container
x86 g
(applicable only g
CXL HCCS
to large
RoCE/NVMe
organizations)
FC SAS IP
IB Quantum dot
PCM MRAM
HDD SSD

Biomolecule Optical storage

43 Huawei Confidential
Optical Storage Technology

Blu-ray storage Gold nanostructures on glass

Low power Long-term

Long High 100 GB+ per 10 TB per
consumptio stable
service life reliability disk disk
n storage

50 years >>> 600

years
44 Huawei Confidential
DNA Data Storage

A small number of synthetic DNA molecules can store a large amount of
data, and can freeze, dry, transport, and store data for thousands of
years.

Advantages of using DNA as storage media:

Small size

High density

High stability

Bottlenecks and limitations:

High costs of DNA molecular synthesis

Slow data read and search
45 Huawei Confidential
Atomic Storage

In 1959, physicist Richard Feynman suggested that it was possible to use atoms to store
information if they could be arranged the way we wanted.

Because an atom is so small, the capacity of atomic storage will be much larger than
that of the existing storage medium in the same size.

Now, arranging the atoms the way we want has become a reality.

Bottlenecks and limitations:

Strict requirements on the operating environment.

46 Huawei Confidential
Quantum Storage

In our current technology, information in electronic devices is stored and moved through
the flow of electrons.

If electrons are replaced by photons, it may be possible for information to move at the
speed of light.

Although quantum storage has obvious benefits to efficiency and service life, it is
difficult for large-scale deployment.

Quantum storage cannot meet the following requirements:

High storage efficiency

Low noise

Long service life

Operating at room temperature

47 Huawei Confidential
Storage Network Trend
AS-IS TO-BE
FC SAN and IP SAN Converged AI Fabric network
High network costs: The FC private network has low latency but Reduced network costs: Open Ethernet carries high-
high costs. The IP SAN has low costs but high latency and poor performance, low-latency, and low-cost storage networks.
performance. Reduced O&M costs: No dedicated O&M skills are required, and
High O&M costs: IP SAN and FC SAN require dedicated O&M unified network management is supported for data centers.
personnel separately, and do not support cloud-and-network
synergy.

LA Flexibility
N
Flexibility
Eth
No packet loss
NIC NIC
FC FC Low latency
SAN SAN
IP SAN IP SAN
10G Eth
IP FC iNIC
SAN 16G FC SAN iNIC
SAN
10G IP SAN
Low cost Low latency
25/100G Eth

48 Huawei Confidential
Milestones of Huawei Storage Products
201
201 9
First high-end
201 7 Huawei storage
201 25
nd
-gen high- OceanStor
201 accelerated by
3st-gen high- end storage Dorado, fastest AI chips
200 1
1
Launched self- all-flash storage
200 end storage
200 8
Launched 2 -
nd developed
Released
6 SmartMatrix Interop Best
2 into
Dived gen FC HSSDs, pioneer architecture Best-in-class of Show
self-
storage storage. in the all-flash of the year Award
developed
technical IP SAN Founded industry
research Huawei
Symantec Co.,
Ltd.
Start out Grow Lead

Released Huawei
Launched self- Distributed
FC storage Launched T OceanStor
developed clustered
series V3
200 virtual NAS storage
3
200
storage unified storage First SAN&NAS Leader in Industry's first
7 200 converged Gartner Magic full-series
201 storage Huawei
9 Quadrant for the NVMe all-flash
201 OceanStor
2 first time storage
201 Pacific series
4
6 201
202
8
0

49 Huawei Confidential
Architecture of Huawei Intelligent Data Storage Products
Storage Database Big data Edge Intelligent data reduction
• AI-based prediction
Intelligent data Intelligent data storage • Intelligent CPU and GPU reduction
reduction algorithms
• Deep • Intelligent • CPU/GPU acceler- • Intelligent • Intelligent • Intelligent
compression global pre- ation tiering prefetch prediction
Intelligenc diction Intelligent data storage
e • Optimal media combination using
Intelligent scheduling
• Container • Heterogeneous computing • Near-data computing technologies such as intelligent prefetch,
+ power data tiering, hotspot identification, and data
caching
CPU
DRA NPU
SCM GPU
SSD

Intelligent scheduling
Memory • Dynamic management of heterogeneous
Fabric
Network
computing resources; near-data computing
Hardware Network
chip scheduling; quick start of containers
chip
Memory Fabric
All • High-performance network centered on
IP memory
Memory SSD
media media • High-performance network with ns-level
latency
• Memory pooling and tiering
DRAM SC SLC/ QLC/OLC
ns M
µs
QLC
50 µs 200 µs
All IP
• Extremely low cost of SSD media
Edge Production Analysis Backup Archive • In-depth disk-controller collaboration,
SLC/QLC/OLC evolution
• All IP to build an ultra-simplified network

50 Huawei Confidential
 Quiz
1. (Multiple-answer question) Which of the following are data types?
A. Structured data

B. Semi-structured data

C. Unstructured data

D. Massive amounts of data

2. (Multiple-answer question) Which of the following statements about storage are correct?
A. Storage refers to disks.

B. A storage system consists of storage hardware, software, and solutions.

C. Storage types include block storage, file storage, and object storage.

D. File storage is used to store data of applications.

51 Huawei Confidential
 Quiz
3. (Multiple-answer question) Which are characteristics of cloud storage?
A. Convergence

B. Open

C. Elasticity

D. Scale-up

4. (Multiple-answer question) Which of the following are the objectives of integrating AI into
storage?
A. Simple

B. Efficient

C. High power consumption

D. Easy to use

52 Huawei Confidential
Summary

Data and Information

Data Storage Storage Architecture

Trend of Storage
Technologies
Storage Technologies Storage Media

Storage Products Interface Protocols

53 Huawei Confidential
More Information

Enterprise Huawei
Technical Enterprise
Support App Service App

54 Huawei Confidential
Recommendations


Huawei official websites

Enterprise business: [Link]

Technical support: [Link]

Online learning: [Link]

Popular tools

HedEx Lite

Network Document Tool Center

Information Query Assistant

55 Huawei Confidential
Thank you. 把数字世界带入每个人、每个家庭、
每个组织，构建万物互联的智能世界。
Bring digital to every person, home, and
organization for a fully connected,
intelligent world.

Copyright©2020 Huawei Technologies Co., Ltd.

All Rights Reserved.

The information in this document may contain predictive

statements including, without limitation, statements regarding
the future financial and operating results, future product
portfolio, new technology, etc. There are a number of factors that
could cause actual results and developments to differ materially
from those expressed or implied in the predictive statements.
Therefore, such information is provided for reference purpose
only and constitutes neither an offer nor an acceptance. Huawei
may change the information at any time without notice.