Cloud Computing Study Notes - Chapter 4 Storage Virtualization

1. Introduction to storage virtualization

1. What is storage virtualization

  Storage virtualization is to gather multiple storage media (such as hard disks and RAID) through certain technologies to form a storage pool and conduct unified management. From the perspective of the host or workstation, what you see is not multiple hard disks, but partitions or volumes, just like a large-capacity hard disk. This kind of storage system that can manage multiple storage devices in a unified way and provide users with large capacity and high data transmission performance is called virtual storage . In general, storage virtualization is a technology that virtualizes one storage device into multiple storage devices, or virtualizes multiple storage devices into one storage device.

2. The role of storage virtualization

(1) Improve the utilization efficiency of hardware resources (utilize previously unused resources).
(2) Simplify the complexity of system management.
(3) Greatly enhance the reliability of the cloud storage platform to achieve uninterrupted business operation.
(4) For administrators or service providers, storage virtualization can easily adjust storage resources and improve the utilization of storage resources.
(5) For users, centralized storage devices can provide better performance and convenience.

3. Concept of cloud computing storage model

(1) Storage resources: Indicates actual physical storage devices, such as direct-attached storage DAS, network-attached storage NAS, storage area network SAN, etc.
(2) Storage device: Indicates the management unit in storage resources, such as local disk, logical unit number LUN, Storage storage pool, NAS shared directory, etc.
(3) Data storage: Indicates the manageable storage logic unit in the virtualization platform, which carries the virtual machine business, such as creating disks.

2. The process of using virtual storage

(1) Add storage resources (such as SAN, DAS, etc.) on the host software interface, and configure the startup of the host.
(2) After the host associates storage resources, scan storage devices (such as local disks, LUNs, etc.), and scan specific devices to the host.
(3) When the host selects a storage device, add data storage and perform virtualization.
(4) Finally, create volumes and other operations on the virtualized data storage.

2. Common storage types

1. Local Disk

(1) The local disk in the cloud computing virtualization scenario refers to the use of the server's local disk resources, which are provided to the virtualization platform after disk array RAID.
(2) Advantages and disadvantages of using local disk
  ①Easy to use.
  ② No shared framework.
  ③ There is no backup and redundancy mechanism for cross-server weapons.

2. Direct-attached storage DAS

(1) Direct-attached storage, or DAS for short, has a history of 40 years and is a structure in which storage devices are directly connected to servers using storage space. Direct-attached storage DAS provides block-level storage services for servers.
(2) Advantages and disadvantages of using direct-attached storage
  ①Multiple disks are combined into one logical disk to meet the requirements of "mass storage".
  ② It can realize the separation of application data and operating system.
  ③ It can improve access performance.
  ④ Simple implementation.
  ⑤ The server fails and the data is inaccessible.

3. Network Attached Storage NAS

(1) Network Attached Storage, referred to as NAS , is a technology that integrates distributed and independent data and centralizes management so as to access different hosts and application servers. NAS connects storage devices to existing networks to provide data and file services.
(2) Advantages and disadvantages of using network-attached storage NAS
  ①Support advanced features such as snapshots.
  ② Centralized storage.
  ③Provide a secure integrated environment (user authentication and authorization).
  ④The transmission rate is low.
  ⑤ The initial installation cost is high, and the equipment cost is high.

4. Storage Area Network SAN

(1) Storage area network SAN is a high-speed network dedicated to storage operations, usually independent of the computer local area network (LAN). Provides data transmission between the host and the storage system, and the efficiency of data transmission within the network is fast. (Common architectures include FC SAN and IP SAN)
(2) Advantages and disadvantages of using storage area network SAN
  ① High utilization rate of storage capacity.
  ② High compatibility.
  ③The transmission distance is long.
  ④ High bandwidth.
  ⑤ Host and storage devices can be expanded independently.
  ⑥ High cost but complicated.

3. Storage mode

1. Non-virtualized storage

(1) Non-virtualized storage is a traditional storage mode, that is, the disk is partitioned into different logical volumes, and each logical volume can be used by a virtual machine.
(2) Features:
  ① Good performance (there is no intermediate virtualization layer, and the virtual machine reads and writes directly on the disk), fast and efficient.
  ②Few storage functions are supported (snapshots, thin provisioning, etc. are not supported).

2. Virtualized storage

(1) Virtualized storage is to format different storage devices and disks. The purpose of formatting is to shield the differences in the capabilities and interface protocols of the underlying storage devices, and convert various storage resources into unified management data storage resources.
(2) Features
  ①Support multiple storage functions (snapshot, thin disk, disk expansion, storage migration, etc.).
  ②The performance is not high (not as good as non-virtualized storage).

3. Raw device mapping

(1) Raw device mapping is to directly use the hard disk to the virtual machine, so that the virtual machine can directly process the command to call the storage (directly access the disk) in the middle virtualization layer, without any intervention (no volumes need to be created).
(2) Features:
  ① Fast speed (the fastest among the three modes) and good performance.
  ②It supports few storage functions (snapshots, thin provisioning, etc. are not supported), and only supports the use of virtual machines with some operating systems, and data storage can only be used as a whole disk mapped to a raw device and cannot be divided.

4. Storage virtualization method

1. Host-based storage virtualization

(1) If only a single host server (or a single cluster) accesses multiple disk arrays, host-based storage virtualization technology can be used. The work of virtualization is completed on the host server through specific software, and the virtualized storage space can span multiple heterogeneous disk arrays. The biggest advantage is its proven stability and the openness of multi-heterogeneous storage systems.
(2) Features:
  ①The software runs on the host, and the application overhead is relatively large.
  ②Good stability and high compatibility.

2. Device-based storage virtualization

(1) If multiple host servers need to access the same disk array, virtualization technology based on storage devices can be used. At this time, the virtualization work is completed on the array controller, and the storage capacity on an array is divided into multiple storage space LUNs for access by different host systems. This technology is mainly used in the same storage device for data protection and data migration. Its advantages are that it has nothing to do with the host, does not occupy host resources, and has rich data management functions.
(2) The software runs on a dedicated embedded system in the storage device.
(3) Create virtual volumes from storage connected to the SAN.

3. Network-based storage virtualization

(1) Network-based storage virtualization is realized by adding a virtualization engine to the storage area network (SAN), and is mainly used for the integration and unified data management of heterogeneous storage systems. The advantage is that it has nothing to do with the host and does not occupy host resources; it supports heterogeneous hosts and heterogeneous storage devices; it can unify the data management functions of different storage devices; it has a unified management platform and good scalability.
(2) The software runs on the disk array in the storage network.
(3) Create a virtual volume on the storage connected to the disk array.

5. Storage virtualization function

1. Thin Disk and Space Reclamation

(1) Streamlining disks and reclaiming space allows users to improve the efficiency of storage resources and reduce the problem that the unused space of virtual machines occupies too much on the host.
(2) Features:
  ① Allocate as much space as the user uses (automatic allocation).
  ②Space reclamation can release the data space deleted by the user to the data storage.

2. Snapshot

(1) A fully available copy of the specified data set, which includes the image of the corresponding data at a certain point in time (the point in time when the copy starts). A snapshot can be either a copy of the data it represents, or a replica of the data. The function of the snapshot is mainly to enable online data backup and recovery.
(2) Features:
  ① Records the content and status of the virtual machine at a certain point in time.
  ② Restoring virtual machine snapshots can be used to quickly restore the virtual machine to a certain point in time multiple times.
  ③The snapshot includes disk content, virtual machine configuration information, and memory data.
  ④Save differential data between multiple snapshots to save storage space.
(3) When creating a snapshot, a new differential volume will be generated, and the virtual machine will mount this differential volume as a disk file.
(4) Snapshot mode: ROW redirection on write; COW copy on write; WA random write.
(5) Snapshot chain: perform multiple snapshot operations on the virtual machine, and these multiple snapshot operations form a snapshot chain.
(6) The virtual machine volume is always mounted at the end of the snapshot chain.

3. Linked clones

  Map the combination of the source volume and the differential volume as a linked clone volume for use by the virtual machine. A linked clone needs to share a unified virtual disk file with the original virtual machine. Using shared disk files shortens the time to create cloned virtual machines while saving physical disk space. With linked cloning, you can easily create an independent virtual machine for different tasks.

4. Virtual machine virtual disk file migration

  Migrate a virtual machine's disks from one datastore to another. All disks of the virtual machine can be migrated as a whole, or individual disks can be migrated separately. Snapshots of virtual machines can be migrated together. Migration is possible when the virtual machine is turned on or off.

Guess you like

Origin blog.csdn.net/YSL_Lsy_/article/details/126123725