Electrical computers and digital processing systems: multicomput – Distributed data processing – Client/server
Reexamination Certificate
1999-03-01
2003-04-15
Sheikh, Ayaz (Department: 2155)
Electrical computers and digital processing systems: multicomput
Distributed data processing
Client/server
C709S217000, C709S223000
Reexamination Certificate
active
06549934
ABSTRACT:
TECHNICAL FIELD
The invention relates generally to computer systems, and more particularly to a method and system for using computer devices across a computer network or networks.
BACKGROUND OF THE INVENTION
Redundant storage for storing important data is known to provide benefits versus storing such data on a single device. For example, in computer clustering applications, where independent servers are grouped together and managed as a single system for purposes of reliability, redundant storage increases cluster availability since the cluster can continue to access the data if one of the plurality of disks fails. In general, with redundant disk storage, sufficient data is stored on two or more disk drive devices to enable reconstruction of the data in the event of the failure of one disk drive device. Typically the two or more disk drive devices are either co-located or are connected via a single storage bus such as a SCSI (Small Computer System Interface) cable.
However, the use of SCSI technology has a number of drawbacks, a first of which is that SCSI cabling is relatively expensive and hard to configure properly. Perhaps more importantly, SCSI limits the number of devices that can be connected thereto, and also limits the distance that the devices may be separated from one another, since SCSI cable lengths are relatively limited. These are significant problems in disaster recovery applications, where enterprises attempt to redundantly store their data in far-apart locations, e.g., so that a fire in one data center does not disrupt the entire enterprise. As an alternative to SCSI, Fibre channel technology is being used. However, although Fibre channel has increased the length somewhat, it is still relatively limited in length and in the number of devices that can be connected thereto.
Prior redundant storage systems such as remote disk mirroring mechanisms have used device-specific protocols (e.g., a block transfer protocol) to transfer the data to the remote disk drive. However, such protocols are neither flexible nor extensible. For example, if the remote disk drive needs to be changed, the protocol and/or other information may need to be changed on the local machine to handle the new drive. As amounts of data increase, and as disk drives similarly grow in size, such changes are becoming frequent. Moreover, in addition to storage devices, the sharing of other devices such as parallel and serial ports and modems may benefit from remote access. However, no technology is heretofore known to be available that provides access to any type of remote device in a flexible and extensible manner.
SUMMARY OF THE INVENTION
Briefly, the present invention provides a method and system for using the existing structure of an operating system to enable access to and control of many types of remote devices, including disk drives, across a network. A client driver intercepts and queues input/output requests, such as I/O Request Packets (IRPs), that are destined for a target device which an application essentially considers a local device. For example, the target device may include a remote disk device used for purposes of redundant storage. The client driver manages preallocated buffers on the server, and may classify the I/O request to correspond to classes of the buffers. When a buffer is available, the client driver dequeues the I/O request, marshalls it with header information and data, and sends it as a message to the server over one of possibly several connections to the server.
When the message is received at the server, a server driver unmarshalls the message, places its data in one of the preallocated buffers as designated by the client, and generates an I/O request therefrom directed to the server target device. The device performs the requested function and responds with a completion status. The server returns the completion status and data (if any) to the client. The client places the data (if any) in the appropriate location and completes the I/O request with the I/O status of the operation as determined on the server by the target device.
The method and system also may handle cancel operations on the remote device. Further, sequential processing may be ensured, and the client may also load balance across the multiple paths, by selecting a given connection path based on criteria including pending message size and dynamic performance measurements of the connections.
REFERENCES:
patent: 5566331 (1996-10-01), Irwin, Jr. et al.
patent: 5799141 (1998-08-01), Galipeau et al.
patent: 5978815 (1999-11-01), Cabrera et al.
patent: 6014695 (2000-01-01), Yamashita et al.
patent: 6101558 (2000-08-01), Utsunomiya et al.
patent: 6131126 (2000-10-01), Kougiouris et al.
patent: 6134585 (2000-10-01), Yamato
patent: 6219693 (2001-04-01), Napolitano et al.
patent: 6249822 (2001-06-01), Kays, Jr. et al.
Gamache Rod
Kusters Norbert P.
Massa Michael T.
Peterson Glenn R.
Law Offices of Albert S. Michalik PLLC
Microsoft Corporation
Sheikh Ayaz
Tran Philip B.
LandOfFree
Method and system for remote access to computer devices via... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for remote access to computer devices via..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for remote access to computer devices via... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3074458