CN103347082A - Storage framework virtualizing system - Google Patents

Abstract

The invention provides a storage framework virtualizing system which comprises a file system and a storage system, wherein the file system and the storage system are both connected with a local area network through Ethernet switches, the file system is further connected with a plurality of client sides through another local area network, the file system comprises a file server used for block-level volume management and metadata saving and a file server used for file service, and the storage system comprises a storage resource pool which is formed after virtualized storage of several storage devices is provided outwards uniformly. Compared with the prior art, the storage framework virtualizing system has the advantages that virtulization of more than two kinds of storage frameworks is achieved, expansibility and flexibility of the storage system are strengthened finally, cost is reduced, maintenance is improved, practicality is strong, and popularization is easy to achieve.

Description

A Virtualization System of Storage Architecture

technical field

The invention relates to the technical field of computer storage virtualization, and specifically relates to a virtualization system of a storage framework that uses external file metadata and block data to be separately processed, externally integrates disk arrays, and finally realizes SAN storage virtualization.

Background technique

SAN storage has many advantages, such as high performance, high reliability, high scalability, and easy management and maintenance. At the same time, there are also many shortcomings in the actual production environment, mainly reflected in the relationship between storage devices and clients. The interoperability is poor, the construction cost is high, and data files cannot be effectively shared. In the NAS storage structure, the file system is closely connected with the storage subsystem, and the two are integrated, which is also very inconvenient to use. Based on the structural characteristics of NAS and SAN, it is very necessary to design a system that separates the NAS storage structure and realizes the virtualization of SAN storage, and is compatible with the advantages of both.

Contents of the invention

The technical task of the present invention is to solve the deficiencies of the prior art, and provide a virtualization system that uses external file metadata and block data to be separately processed, externally integrates disk arrays, and finally realizes storage architecture of SAN storage virtualization.

The technical scheme of the present invention is realized in the following manner. The virtualization system of this storage framework includes a file system and a storage system connected to a local area network through an Ethernet switch, wherein the file system is also connected to several The client, the file system includes a file server that can be used for block-level volume management and metadata storage, and a file server that can be used for file services, and the storage system includes a storage resource pool, which is composed of several storage devices The storage after virtualization is uniformly provided externally.

The storage resource pool provides storage space externally and manages storage internally. When the client performs IO operations such as reading and writing data on the storage space, the I/O command is first sent to a dedicated metadata server. The data server first obtains the metadata and the view of the stored data, and the dedicated metadata server obtains the required data through the storage framework according to the metadata and the data view, and then transmits the data to the client to meet the IO request. The data server refers to a metadata server that stores all data views of the entire backend virtual pool and handles metadata access.

The virtualization implementation of the storage architecture consists of two parts, one part is located in the metadata server, which is responsible for the virtualization of the back-end storage; the other part is located in the client part, and redirects the IO request: when the client's IO request is not locally When storing, the virtualization mechanism sends the IO command to the metadata server, and the metadata server returns the corresponding data view and metadata part to the client after receiving the IO request.

The beneficial effect that the present invention produces compared with prior art is:

The virtualization system of a storage framework of the present invention integrates the technical features of SAN and NAS, realizes the storage virtualization of the disk array, guarantees many advantages of the SAN, and realizes the virtualization management of the disk array at the same time. This design separates the file system and the storage subsystem in the traditional NAS structure, and after separating the two, it is composed of a server respectively. The main task of the server with the file system function is to save the metadata and execute the function of the file system, etc.; The main function of the server carrying the storage subsystem is to complete the advanced functions of the storage array; the above-mentioned structure is adopted to realize the virtualization of more than two storage architectures. Finally, the scalability and flexibility of the storage system are enhanced, and the cost and maintenance are reduced and improved to a certain extent. It is practical and easy to popularize.

Description of drawings

Accompanying drawing 1 is a structural representation of the present invention.

Detailed ways

A storage architecture virtualization system of the present invention will be described in detail below in conjunction with the accompanying drawings.

As shown in Figure 1, a virtualization system of a storage framework includes a file system and a storage system connected to a local area network through an Ethernet switch, wherein the file system is also connected to several clients through another local area network. The system includes a file server that can be used for volume management at the block level and saves metadata, and a file server that can be used for file services. The storage system includes a storage resource pool, which is a storage resource pool that is virtualized by several storage devices Formed by unified external provision.

The storage resource pool provides storage space externally and manages storage internally. When the client performs IO operations such as reading and writing data on the storage space, the I/O command is first sent to a dedicated metadata server. The data server first obtains the metadata and the view of the stored data, and the dedicated metadata server obtains the required data through the storage framework according to the metadata and the data view, and then transmits the data to the client to meet the IO request. The data server refers to a metadata server that stores all data views of the entire backend virtual pool and handles metadata access.

The virtualization implementation of the storage architecture consists of two parts, one part is located in the metadata server, which is responsible for the virtualization of back-end storage; the other part is located in the client part, and redirects the IO request: when the client's IO request is not locally When storing, the virtualization mechanism sends IO commands to the metadata server, and the metadata server returns the corresponding data view and metadata part to the client after receiving the IO request.

In this design, a separate server is added at the SAN level to provide virtualization services. Through the virtualization function, various storage devices at the back end of the SAN are virtualized, and the virtualized storage is uniformly provided externally. Formed to store resource pools.

In order to achieve unified storage, complete the further integration of NAS and SAN. The use of two dedicated servers, a metadata server and a file server, the former is used for block-level volume management, and the latter is used for file services. Based on this architecture, the system can provide both File IO and BLOCK IO services. On the storage side, combined with the integration of NAS and SAN structures, the storage system can simultaneously provide two storage parts: File space for file services and BLOCK space for data block services.

Except for the technical features described in the description, all are well-known technologies of those skilled in the art.

Claims (3)

1. A virtualization system of a storage framework, characterized in that: its structure includes a file system and a storage system all connected to a local area network through an Ethernet switch, wherein the file system is also connected to several clients through another local area network, and the file system It includes a file server that can be used for block-level volume management and saves metadata, and a file server that can be used for file services. The storage system includes a storage resource pool, which is a unified storage after virtualization of several storage devices. Provided externally. 2. The virtualization system of a storage framework according to claim 1, characterized in that: the storage resource pool provides external storage space, internally performs unified storage management, and the client performs data storage for the storage space During IO operations such as reading and writing, the I/O command is first sent to the dedicated metadata server, and the metadata and the view of the stored data are first obtained on the metadata server. The dedicated metadata server obtains the metadata and data view through the storage architecture. The required data is then transmitted to the client to meet the IO request. The above-mentioned dedicated metadata server refers to a metadata server that stores all data views of the entire back-end virtual pool and handles metadata access. 3. The virtualization system of a kind of storage framework according to claim 1, it is characterized in that: the virtualization realization of described storage framework is made up of two parts, and a part is positioned at metadata server, is responsible for the virtualization of back-end storage; The other part is located in the client part, which is the redirection operation of IO requests: when the client's IO request is not stored locally, the virtualization mechanism sends the IO command to the metadata server, and the metadata server sends the corresponding data view after receiving the IO request and the metadata part are returned to the client.

Images