Electrical computers and digital processing systems: multicomput – Computer-to-computer session/connection establishing – Network resources access controlling
Reexamination Certificate
1999-03-03
2002-09-17
Le, N. (Department: 2858)
Electrical computers and digital processing systems: multicomput
Computer-to-computer session/connection establishing
Network resources access controlling
C709S213000
Reexamination Certificate
active
06453354
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates generally to data storage systems, and more particularly to network file servers.
2. Background Art
Mainframe data processing, and more recently distributed computing, have required increasingly large amounts of data storage. This data storage is most economically provided by an array of low-cost disk drives integrated with a large semiconductor cache memory. Such cached disk arrays were originally introduced for use with IBM host computers. A channel director in the cached disk array executed channel commands received over a channel from the host computer.
More recently, the cached disk array has been interfaced to a data network via at least one data mover computer. The data mover computer receives data access commands from clients in the data network in accordance with a network file access protocol such as the Network File System (NFS). (NFS is described, for example, in RFC 1094, Sun Microsystems, Inc., “NFS: Network File Systems Protocol Specification,” Mar. 1, 1989.) The data mover computer performs file locking management and mapping of the network files to logical block addresses of storage in the cached disk storage subsystem, and moves data between the client and the storage in the cached disk storage subsystem.
In relatively large networks, it is desirable to have multiple data mover computers that access one or more cached disk storage subsystems. Each data mover computer provides at least one network port for servicing client requests. Each data mover computer is relatively inexpensive compared to a cached disk storage subsystem. Therefore, multiple data movers can be added easily until the cached disk storage subsystem becomes a bottleneck to data access. If additional storage capacity or performance is needed, an additional cached disk storage subsystem can be added. Such a storage system is described in Vishlitzky et al. U.S. Pat. No. 5,737,747 issued Apr. 7, 1998, entitled “Prefetching to Service Multiple Video Streams from an Integrated Cached Disk Array,” incorporated herein by reference.
Unfortunately, data consistency problems may arise if concurrent client access to a read/write file is permitted through more than one data mover. These data consistency problems can be solved in a number of ways. For example, as described in Vahalia et al., U.S. Pat. No. 5,893,140 issued Apr. 6, 1999, entitled “File Server Having a File System Cache and Protocol for Truly Safe Asynchronous Writes,” incorporated herein by reference, locking information can be stored in the cached disk array, or cached in the data mover computers if a cache coherency scheme is used to maintain consistent locking data in the caches of the data mover computers. However, as shown in
FIG. 1
, labeled “Prior Art,” a more elegant solution to the data consistency problem has been implemented at EMC Corporation in a network file server system having multiple stream server computers and one or more cached disk arrays.
FIG. 1
shows a network file server system having at least two data mover computers
21
and
22
. The first data mover
21
has exclusive access to read/write files in a first file system
23
, and the second data mover
22
has exclusive access to read/write files in a second file system
24
. As shown, the file systems
23
,
24
are respective volumes of data contained in the same cached disk array
25
, although alternatively each file system
23
,
24
could be contained in a respective one of two separate cached disk arrays. For example, each of the data mover computers
21
,
22
has a respective high-speed data link to a respective port of the cached disk array
25
. The cached disk array
25
is configured so that the file system
23
is accessible only through the data port connected to the first data mover
21
and so that the file system
24
is accessible only through the data port connected to second data mover
22
. Each of the data movers
21
,
22
maintains a directory of the data mover ownership of all of the files in the first and second file systems
23
,
24
. In other words, each of the data movers maintains a copy of the file system configuration information in order to recognize which data mover in the system has exclusive access to a specified read/write file.
Each of the data movers
21
,
22
may receive file access requests from at least one network client. For example, the first data mover
21
has a network port
28
for receiving file access requests from a first client
26
, and the second data mover
22
has a network port
29
for receiving file access requests from a second client
27
. The clients
26
,
27
communicate with the data movers using the connection-oriented NFS protocol. Whenever the data mover
21
receives a file access request from the client
26
, it checks the configuration directory to determine whether or not the file specified by the request is in a file system owned by the data mover
21
. If so, then the data mover
21
places a lock on the specified file, accesses the file in the file system
23
, and streams any read/write data between the client
26
and the file system
23
. If the file specified by the request is not a file system owned by the data mover
21
, then the data mover
21
forwards the request to the data mover that owns the file system to be accessed. For example, if the client
26
requests access to a file in the file system
24
, then the first data mover
21
forwards the file access request to the second data mover
22
. The second data mover
22
places a lock on the file to be accessed, the second data mover accesses the file, and the second data mover streams any read/write data between the first data mover
21
and the file in the file system
24
. The first data mover then streams the read/write data between the second data mover
22
and the client
26
. The second data mover
22
responds to file access requests from its client
27
in a similar fashion, by directly servicing file access request to files in the file system
24
that it owns, or forwarding to other data movers the requests for access to the files in file systems that it does not own.
The solution as shown in
FIG. 1
is rather efficient because the data movers
21
,
22
can be linked by a dedicated high-speed data link for the exchange of read/write data between them. Therefore, there is no additional loading of the data network between the data movers and the clients and no additional loading of the data links between the cached disk array
25
and the data movers
21
,
22
. The data movers can cache the file access information (e.g., file locks) and file data and attributes for the files that they own, so that the loading on the data links between the cached disk array and the data movers
21
,
22
can be somewhat reduced. In the network file system implemented at EMC Corporation, when a data mover did not own the file system to be accessed, the data mover forwarded to or exchanged NFS data packets with the data mover that owned the file system to be accessed. Such a system was relatively easy to implement, since it involved creating a proxy router routine that would recognize whether or not a NFS data packet from a client was for access to a file system owned by another data mover, and if so, routing the data packet to the data mover that owned the file system. The data mover owning the file system could treat the forwarded data packet in a fashion similar to a data packet received directly from a client.
Although the system of
FIG. 1
is satisfactory for handling NFS file access requests, it has a number of limitations that will become increasingly significant. The current trend is toward higher-speed network links and interconnection technology, such as technology for the Fibre-Channel standards being developed by the American National Standards Institute (ANSI). In a network employing high-speed links and interconnection technology, the delays inherent in a connectionless communications protocol such as NFS become more pronounced.
The I
Gupta Uday
Jiang Xiaoye
Tzelnic Percy
Vahalia Uresh K.
Benson Walter
EMC Corporation
Howrey Simon Arnold & White , LLP
Le N.
LandOfFree
File server system using connection-oriented protocol and... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with File server system using connection-oriented protocol and..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and File server system using connection-oriented protocol and... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2860743