stornext

A shared-disk file system uses a storage-area network (SAN) to provide direct disk access from multiple computers at the block level.（采用块级访问） Access control and translation from file-level operations that applications use to block-level operations used by the SAN must take place on the client node. （在客户端进行访问控制和文件级操作-块级操作的转换）The most common type of clustered file systems is shared-disk file system, （最普遍的cluster file system就是shared disk file system）which—by adding mechanisms for concurrency control—provides a consistent and serializable view of the file system, avoiding corruption and unintended data loss even when multiple clients try to access the same files at the same time. It is a common practice for shared-disk filesystems to employ some sort of a fencing mechanism to prevent data corruption in case of node failures, （能够保证同时访问，是采用了一定的隔离机制）because an unfenced device can cause data corruption if it loses communication with its sister nodes, and tries to access the same information other nodes are accessing.（这种隔离机制以姐妹节点之间的通讯和可以访问同样的信息为前提）

The underlying storage area network may use any of a number of block-level protocols, including SCSI, iSCSI, HyperSCSI, ATA over Ethernet (AoE), Fibre Channel, network block device, and InfiniBand.（SAN所采用的块级协议）

There are different architectural approaches to a shared-disk filesystem. Some distribute file information across all the servers in a cluster (fully distributed). （shared-disk filesystem有不同的架构，有些把文件信息完全分散到各个节点（组成集群），有些就是采用集中式的MDC）Others utilize a centralized metadata server. Both achieve the same result of enabling all servers to access all the data on a shared storage device.

A common performance measurement of a clustered file system is the amount of time needed to satisfy service requests. In conventional systems, this time consists of a disk-access time and a small amount of CPU-processing time. But in a clustered file system, a remote access has additional overhead due to the distributed structure. This includes the time to deliver the request to a server, the time to deliver the response to the client, and for each direction, a CPU overhead of running the communication protocol software.（cluster file system对性能的测试就是需要满足对服务请求的响应时间。在传统的文件系统中，只需考虑磁盘访问和cpu处理的时间，但是在cluster file system中，由于分布式，所以还得考虑服务器和客户端的相互通讯的响应时间，以及cpu对运行通讯协议软件的开销）

Examples[edit]

• Silicon Graphics (SGI) clustered file system (CXFS)

• Veritas Cluster File System

• DataPlow Nasan File System

• IBM General Parallel File System (GPFS)

• Lustre

• Microsoft Cluster Shared Volumes (CSV)

• Oracle Cluster File System (OCFS)

• PolyServe storage solutions

• Quantum StorNext File System (SNFS), ex ADIC, ex CentraVision File System (CVFS)

• Blue Whale Clustered file system (BWFS)

• Red Hat Global File System (GFS)

• Sun QFS

• TerraScale Technologies TerraFS

• Versity VSM

• VMware VMFS

• Xsan

——————————————————————————————————————————

对比：

Distributed file systems[edit]

Distributed file systems do not share block level access to the same storage but use a network protocol.^{[2][3]（分布式文件系统采用网络协议访问）} These are commonly known as network file systems, （）even though they are not the only file systems that use the network to send data.^{[citation needed]} Distributed file systems can restrict access to the file system depending onaccess lists or capabilities on both the servers and the clients, （分布式文件系统可以在客户端或者服务器端，根据访问列表限制对文件系统的访问）depending on how the protocol is designed.

The difference between a distributed file system and a distributed data store is that a distributed file system allows files to be accessed using the same interfaces and semantics as local files – for example, mounting/unmounting, listing directories, read/write at byte boundaries, system's native permission model. Distributed data stores, by contrast, require using a different API or library and have different semantics (most often those of a database).使用分布式文件系统，可以让别的客户端像本地文件系统一样操作。

A distributed file system may also be created by software implementing IBM's Distributed Data Management Architecture (DDM), in which programs running on one computer use local interfaces and semantics to create, manage and access files located on other networked computers. All such client requests are trapped and converted to equivalent messages defined by the DDM. Using protocols also defined by the DDM, these messages are transmitted to the specified remote computer on which a DDM server program interprets the messages and uses the file system interfaces of that computer to locate and interact with the specified file.

分布式文件系统对于客户端来讲，希望是透明的，对于客户端来讲，就像本地文件系统一样。

访问透明

位置透明，远程文件和本地文件在同一个命名空间内，远程文件不需要提供位置，IP

跨平台复制和传输，不需要关心平台

对于本地和远程来说，做到文件系统的一致性

Examples[edit]

Main article: List of distributed file systems

• BeeGFS (Fraunhofer)

• Ceph (Inktank, Red Hat, SUSE)

• Infinit

• GFS (Google Inc.)

• Windows Distributed File System (DFS) (Microsoft)

• GlusterFS (Red Hat)

• HDFS (Apache Software Foundation)

• LizardFS (Skytechnology)

• MooseFS (Core Technology / Gemius)

• OneFS (EMC Isilon)

• OrangeFS (Clemson University, Omnibond Systems), formerly Parallel Virtual File System

• ObjectiveFS

• Panfs (Panasas)

• RozoFS (Rozo Systems)

• Parallel Virtual File System (Clemson University, Argonne National Laboratory, Ohio Supercomputer Center)

• XtreemFS

————————————————————————————————————————————————

对比：

NAS：

Network-attached storage (NAS) provides both storage and a file system, like a shared disk file system on top of a storage area network (SAN). （基于文件协议）NAS typically uses file-based protocols (as opposed to block-based protocols a SAN would use) such as NFS (popular on UNIX systems), SMB/CIFS (Server Message Block/Common Internet File System) (used with MS Windows systems), AFP (used with Apple Macintosh computers), or NCP (used with OES and Novell NetWare).

————————————————————————————————————————————————

StorNextconsists of two components:

• StorNext FileSystem(SNFS), a high performance data sharing software

• StorNextStorage Manager (SNSM), the intelligent, policy-based data mover，The StorNext Storage Manager is a policy based data management system that can copy, migrate and/or archive data from the StorNext File System into a variety of storage devices in multiple locations. Data can be tiered into disk, a Quantum Lattus object store, a robotic tape library, or even exported into an offline vault. 

distributed LAN clients 

• . Unlike a traditional StorNextSAN client, a distributed LAN client does not connect directly toStorNext via fibre channel or iSCSI, but rather across a LAN through agateway system called adistributed LAN server 

• . The distributed LANserver is itself a directly connected StorNext client, but it processes requests from distributed LAN clients in addition to running applications.

他的lisence，有一定的试用期，lisence包含各种功能，需要选。

Here is a list of StorNext licenses:

•File System: A File System license enables you to create and modify StorNext-supported file systems.

•LAN Client: You must have a Distributed LAN Client license for each LAN client you use with StorNext (in addition to any SAN clients).

•Storage Manager: A Storage Manager license provides full access to StorNext's Storage Manager features that are not licensed separately.

•Replication: A replication license is required if you want to use StorNext's Data Replication feature.

•Deduplication: A deduplication license is required if you want to use StorNext’s Data Deduplication (blockpool) feature.

•Vaulting: A Vaulting license provides the ability to move seldom-used media to a manual archive vault, freeing room for media in the managed archives.

•Storage Disk: You must have a Storage Disk license to be able to configure and use StorNext storage disks.

•Checksum: A Checksum license enables you to verify data integrity by ensuring that the checksum created when data was stored matches the checksum upon data retrieval.

•Distributed Data Mover (DDM): A license is required for using DDM if you plan to use additional machines besides the primary server to distribute data streams.

•Failover (HA): A Failover (High Availability) license is required if you plan to use StorNext's HA failover features.

•Maintenance: A Maintenance license verifies that your site has purchased StorNext upgrade licenses, and is required for StorNext upgrades. It is also used at run time to verify the revision number in the software matches what was purchased

StorNext consists of two components: StorNext FileSystem (SNFS), a high performance data sharing software, and StorNextStorage Manager (SNSM), the intelligent, policy-based data mover