001 Storage Basics and Application Environments
001 Storage Basics and Application Environments
Environments
Contents
1. What Is Storage? 4. What Are the Other Basic Storage Concepts?
• Storage Concepts • RAID
• Storage in Data Centers • LUN
• Storage Evolution • Multipathing
• Block Storage • File System
• File Storage • Local File System
• Object Storage • Network File System
2. What Are the Basic Concepts of Storage Hardware (L2)? • CIFS
• Disks (NVMe SCM) • NFS
• Disk Arrays • Storage Performance Indicators
• Controllers 5. What Are the Basic Concepts of Data Protection?
• Controller Enclosures • RPO and RTO
• Disk Enclosures • Backup
• High-Density Enclosures • Disaster Recovery
• I/O Modules and HBAs • Snapshot
3. What Are the Basic Concepts of Storage Protocols? • Remote Replication
• SCSI • Mirroring
• NVMe • Clone
• iSCSI
• Fibre Channel
• SAS
1 Huawei Confidential
What Is Storage?
— Storage Is the Media for Storing and Protecting Data
Storage in a narrow sense Storage in a broad sense
Server
Multipathin
g software
Fibre
Channel
switch Fibre
Channel
link
Controller
1. Storage hardware (disk arrays, controllers, disk enclosures, and tape libraries)
Disk array
2. Storage software (management software, snapshot, replication, and multipathing software)
3. Storage networks (HBAs, Fibre Channel switches, as well as Fibre Channel and SAS cables)
4. Storage solutions (centralized storage, archiving, backup, and disaster recovery)
2 Huawei Confidential
Storage Concepts
Storage is to save data to certain storage media in a reasonable, secure, and effective manner to meet requirements
of different application environments and ensure effective accesses to the data. Specifically:
1. Storage is physical media for temporary or long-term data storage.
2. Storage is a method or behavior for ensuring data integrity and security. Storage combines the two aspects to
provide customers with a data storage solution.
3 Huawei Confidential
Storage in Data Centers
OceanStor 18500 V5
OceanStor or 18800 V5
OceanStor 6800 V5
OceanStor
5000 series V5
Application 2200 V3 or
(mid-range)
2600 V3
Runtime library
Database
Dorado 8000 V6 or OceanStor FusionStorage
Server 18000 V6 9000
Storage
Network
4 Huawei Confidential
Storage Evolution
External disk array (DAS) Smart disk array (DAS)
SAN NAS
Disks in a server Server
CPU Server CPU Application Application
Multiple
server server
application
Server Arm Arm
File system File system servers
CPU
Limitations JBOD logically connects several A controller provides RAID and large-capacity cache, Resolved issues
• Disks become the system performance physical disks to increase capacity. It enables the disk array to have multiple functions, • Disks become the system performance
bottleneck. does not provide data protection. and is equipped with dedicated management bottleneck.
• The number of disk slots is limited, software. • The number of disk slots is limited,
thereby providing limited capacity. Resolved issues Resolved issues thereby providing limited capacity.
• Data is stored on single disks, and • The number of disk slots is limited, • Disks become the system performance • Data is stored on single disks, and data
data storage is less reliable than thereby providing limited capacity. bottleneck. storage is less reliable than expected.
expected. • Data is stored on single disks, and • The number of disk slots is limited, providing a • Storage space utilization is low.
• Storage space utilization is low. data storage is less reliable than small capacity. • Data is scattered in local storage
• Data is scattered in local storage expected. • Data is stored on single disks, and data storage is systems.
systems. less reliable than expected.
5 Huawei Confidential
File Storage
NFS or CIFS
6 Huawei Confidential
Block Storage
7 Huawei Confidential
Object Storage
HTTP, REST, SOAP, or S3
8 Huawei Confidential
What Are the Basic Concepts of
Storage Hardware (L2)?
9 Huawei Confidential
Disks
10 Huawei Confidential
SSDs VS. HDDs
SSD HDD
Price High: several yuan (RMB) per GB Low: several jiao (RMB) per GB
11 Huawei Confidential
Disk Types
Medi Functions
aHDD Disk types Desktop level
Monitoring level
SSD Enterprise level
12 Huawei Confidential
Disk Types — Serial ATA (SATA)
Advantages
1. SATA is a serial bus interface protocol. During data transmission, the data and signal
cables are used independently, and the embedded clock frequency signal is used. Its rate
can reach 30 times of the Parallel ATA (PATA).
2. Instead of simple PATA improvement, the bus structure is brand new.
3. Control information is scattered in data and transmitted by using a predefined bit.
4. One path is used to transmit data, and the other is used to return a response.
5. SARA has higher anti-interference capabilities and faster speed than PATA. The
installation is easier and the number of cables used in the chassis is reduced.
6. The performance of single-thread tasks is good.
Disadvantages
1. SATA, designed for entry-level applications, is not as powerful as SCSI in terms of big
data throughput or multi-thread transmission.
2. When multiple threads are reading data, the head of a disk swings back and forth,
causing the disk to be overheated.
13 Huawei Confidential
Disk Types — Serial Attached SCSI (SAS)
1. SCSI is a storage interface protocol specially designed for small computer systems. 50-
pin ports are used. The appearance of a 50-pin port is similar to that of a common disk
port. The SCSI disks support connections to various devices. In addition, each SCSI disk
has an independent chip for data processing. The CPU usage is low, the bandwidth can
reach 320 MB/s, and the stability is good.
2. SAS is a disk connection technology that integrates the advantages of the parallel SCSI
and serial connection technologies.
3. SAS is a point-to-point, full-duplex, and dual-port interface.
4. SAS is compatible with SATA, meeting high-performance requirements of enterprises,
achieving interoperability with SATA, and bringing unprecedented flexibility and
benefits to enterprises.
5. SAS features high performance, high reliability, and powerful scalability.
The serial technology is also used, which is better than SCSI in transmission rate and anti-
interference. However, the price is higher.
14 Huawei Confidential
Disk Types — Near-Line SAS (NL-SAS)
1. NL-SAS disks integrate SAS interfaces and SATA disks. In other words, NL-SAS disks
refer to SATA disks with SAS interfaces and near-SAS performance.
2. Near-line storage is oriented to applications between online storage and offline
storage. Data that is not frequently used or seldom accessed is stored on the storage
devices of which performance is relatively low. However, these devices must provide
fast addressing capabilities and a high transmission rate.
15 Huawei Confidential
Disk Types — SSDs
SSD hardware structure
6 Gbit/s Multi-channel
SAS port concurrence
SSD Flash
controller
DDR
17 Huawei Confidential
Basic Concepts Related to RAID
Redundant Array of Inexpensive Disks or Drives, or 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.
RAID is classified into different RAID levels based on Two different RAID levels can be combined to form a
the combination methods. new RAID level.
RAID 0 Known as a stripe set or striped volume, it splits data evenly RAID 0+1 RAID 01, also called RAID 0+1, is a RAID level using a
across two or more disks, without parity information, mirror of stripes, achieving both replication and
redundancy, or fault tolerance. sharing of data between disks.
RAID 1 It consists of an exact copy (or mirror) of a set of data on
two or more disks; a classic RAID 1 mirrored pair contains
RAID 10 RAID 10, also called RAID 1+0 and sometimes RAID
two disks. This configuration offers no parity, striping, or
1&0, is similar to RAID 01 with an exception that two
spanning of disk space across multiple disks, since the data
used standard RAID levels are layered in the opposite
is mirrored on all disks belonging to the array, and the array
order; thus, RAID 10 is a stripe of mirrors.
can only be as big as the smallest member disk.
RAID 3 It is rarely used in practice. It consists of byte-level striping
with a dedicated parity disk.
RAID 5 It consists of block-level striping with distributed parity. RAID 50 RAID 50, also called RAID 5+0, combines the straight
block-level striping of RAID 0 with the distributed
RAID 6 RAID 6 extends RAID 5 by adding another parity block; thus,
parity of RAID 5.
it uses block-level striping with two parity blocks distributed
across all member disks.
18 Huawei Confidential
SCM — Next-Generation Storage Media
Storage Class Memory (SCM) is a new storage High-performance SSDs provide the following
medium popular in the industry. SCM is similar to application forms in a storage system:
storage in its persistence and to memory in its Metadata cache: As metadata cache of AFA,
byte-level access. SCM SSDs work with the DRAM to build a
memory + SCM SSD two-tier cache, which
The SCM SSD, which uses NVMe block interface avoids bottlenecks in memory capacity and
and is compatible with the native architecture, is supports larger user capacity with stable
the primary application form of SCM. Optane performance.
P4800X series launched by Intel is an example.
This product has little impact on the system Data cache: SCM SSDs serve as the acceleration
architecture but provides better performance layer of user data and improves performance in
than that of flash SSDs. In addition, SCM SSDs do typical application scenarios.
not require garbage collection, which prevents
performance deterioration similar to that of Main storage: SCM SSDs serve as the storage
NAND SSD after long-time running, and curbs layer of user data and provides a high-
latency at an appropriate level. This type of performance storage system to meet the
storage products is springing up in the industry, performance requirements of some scenarios.
and Huawei is also working on the R&D of
relevant products.
19 Huawei Confidential
Disk Arrays
20 Huawei Confidential
Disk Array Composition
Separation of the controller
enclosure and disk enclosure
+ =
Integration of controllers
and disk enclosure
+ =
21 Huawei Confidential
Controllers
22 Huawei Confidential
Controller Enclosures
Power module
Control
BBU module
23 Huawei Confidential
Disk Enclosures
3.5-inch disk enclosure
24 x 3.5-inch disks
Support for mainstream SAS
Power modules
2 + 2 redundancy
90% power conversion
efficiency
24 Huawei Confidential
High-Density Enclosures
25 Huawei Confidential
I/O Modules and HBAs
An I/O module connects a storage system to a server or switch for data transmission. The I/O
module includes field pluggable units with several interfaces.
HBA is short for Host Bus Adapter. An HBA is a circuit board that enables a computer to provide
I/O processing and physical connections between a server and a storage device.
26 Huawei Confidential
What Are the Basic Concepts of
Storage Protocols?
27 Huawei Confidential
SCSI
28 Huawei Confidential
iSCSI
29 Huawei Confidential
Fibre Channel
30 Huawei Confidential
SAS
31 Huawei Confidential
NVMe
32 Huawei Confidential
What Are the Other Basic Storage
Concepts?
33 Huawei Confidential
RAID
The term RAID was invented by David Patterson, Garth A. Gibson, and Randy Katz at
the University of California, Berkeley in 1987. It combines multiple independent
physical disks into a virtual logical disk using related algorithms to provide larger
capacity, higher performance, and better error tolerance capabilities.
RAID 0 RAID 5
RAID 1 RAID 6
RAID 10
RAID 50
RAID DP RAID MP
RAID 1E RAID 5EE
RAID 5E RAID ADG
34 Huawei Confidential
LUN
One or multiple logical volumes can be created for RAID based on the specified
capacity. A logical volume is identified by logical unit number (LUN).
RAID 10 RAID 5
35 Huawei Confidential
Multipathing
37 Huawei Confidential
Local File System
File systems and application programs are on a same server.
Storage silos
38 Huawei Confidential
Network File System
File system client File system client File system client File system client
Network protocol
39 Huawei Confidential
CIFS
The Common Internet File System (CIFS) is a mainstream
share file system developed by Microsoft for serving
heterogeneous platforms and is mainly applied in
Windows. Client systems use the TCP or IP protocol to
request file access services from server systems over a
network.
CIFS share authentication provides two types of shared
file access permissions: user and Active Directory Server
(ADS).
The CIFS normal share means that the file system is
shared as a directory and all users can access the
directory.
The CIFS homedir share is a file sharing mode provided by
file engines. The CIFS homedir share only allows a user to
access the directory named with the user name and each
user can only access a directory that belongs to the user's
40directory.
Huawei Confidential
NFS
41 Huawei Confidential
IOPS
42 Huawei Confidential
OPS
43 Huawei Confidential
Performance Indicators and Evaluation Methods of Storage Devices
Performance indicators
IOPS is the number of I/Os that can be processed by a storage device per second. It is used to measure the response
capability of a storage device. IOPS is the most important measurement indicator for a large number of small I/Os.
Bandwidth is also called throughput that indicates the total amount of data that can be processed per second. It is
used to measure the storage throughput. Bandwidth is useful for measuring large I/Os, especially for measuring the
time required for processing a large amount of data.
Bandwidth = IOPS x Average I/O size
Latency refers to the time consumed for processing I/Os. It is used to measure the processing speed of storage
devices. Latency is classified into host latency and storage latency. Storage latency refers to the period from the time
when I/Os arrive at the storage device to the time when the storage device returns a processing completion message
to the host. Host latency is the sum of the storage latency, link transmission time, and host queuing time. Users focus
on the host latency.
IOPS = Number of concurrent requests/Average latency
Performance evaluation
IOPS and bandwidth are the two most important indicators for performance evaluation.
The IOPS assessment focuses on the I/O sequence, cache hit ratio, and IOPS of a single disk. The major bottleneck of
the maximum IOPS of a storage system is the CPU processing capability.
The bandwidth assessment focuses on the I/O size, bandwidth of a single disk, and storage hardware bandwidth. For
the maximum bandwidth of a storage system, the major bottleneck is the front-end and back-end channel bandwidth
and mirroring bandwidth.
44 Huawei Confidential
What Are the Basic Concepts of Data
Protection?
45 Huawei Confidential
Backup
Online application information is extracted to create one or more copies based on
certain policies, and the copies are stored on preset storage media for recovery in case
of an online system fault.
46 Huawei Confidential
Three Elements of a Backup System
RPO
Backup Backup Point in time to An error or Recovery Recovery Application
started completed which data is disaster startup completed recovery
recovered occurs.
12:00
BW + RTO
00:00 = Economic loss
06:00
47 Huawei Confidential
Disaster Recovery
Disaster recovery is a higher-level data protection.
48 Huawei Confidential
Disaster Recovery and Its Tiers
Disaster recovery refers to establishing a systematic data emergency response mode in advance by using scientific
technical means and methods to cope with disasters. The content includes data backup and system backup, business
continuity planning, personnel architecture, communication assurance, crisis management, disaster recovery planning,
disaster recovery schemes, business recovery schemes, emergency response, third-party cooperation organizations, and
supply chain crisis management. The disaster recovery levels range from the module level to system-level and solution-
level.
SHARE's seven tiers of disaster recovery released in 1992, were updated in 2012 by IBM as an eight tier model.
(Generally, data backup is usually applied to tier 1 to tier 4, and data disaster recovery is usually applied to tier 4 to tier 7.)
TCO
Tier 7: highly automated, business-integrated solution
51 Huawei Confidential
Snapshot (HyperSnap)
a b c a b c
08:00 am d e f Snapsh d e f
g h i ot g h i
j k l j k l
a b c a b c
d m f d e f
09:00 am
g h n g h i
j k l j k l
52 Huawei Confidential
Mirroring (HyperMirror)
1. Function: Data is stored concurrently on two independent storage areas
(commonly logical volumes) through the same I/O operations.
2. Benefit
• The two mirror volumes are identified by the host or storage controlled as
the same production volume. Services can be switched over between the
two volumes to improve system reliability.
• Concurrent operations on mirror volumes can improve system performance.
I/O
LUN 1 LUN 2
53 Huawei Confidential
Clone (HyperClone)
A clone is a copy or multiple copies of source data at a specific point in time.
A snapshot is similar to a stereotyped shadow at a certain point in time, while a
clone is a stereotyped entity at a certain point in time.
4
2
1 3
Implementation process
Primary LUN Secondary LUN Primary LUN Secondary LUN
11 22 33 44
Enable the clone function. After data After splitting is The secondary LUN can
Start data synchronization is completed, the primary be mapped to the host. It
synchronization. complete, data on the and secondary LUNs can be used for data
primary LUN is the same become independent analysis and query.
as that on the secondary LUNs.
LUN.
54 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.