System and Method for Achieving Different Levels of Data Consistency

Abstract

A system and method for maintaining consistency in a system where multiple copies of an object may exist is provided for maintaining consistent copies. Consistency is maintained using a plurality of consistency policies in which at least one consistency policy results in different performance than a second consistency policy. A consistency policy is selected from the plurality consistency policies for each object to improve system performance.

Description

RELATED APPLICATION INFORMATION[0001]This application is a Continuation of U.S. patent application Ser. No. 10 / 715,237 filed Nov. 17, 2003, which is incorporated herein by reference in its entirety.BACKGROUND OF THE INVENTION[0002]1. Field of the Invention[0003]The present invention relates to data storage and more particularly to systems and methods for achieving data consistency among multiple copies.[0004]2. Description of the Related Art[0005]Many computer applications create multiple copies of the same data. Maintaining consistency of these multiple copies is critically important. How the updating of the different copies is coordinated leads to different levels of consistency among the copies, in return for different costs to perform that coordination. Typically, a stronger consistency, with closer coordination between peer cache updates, results in a larger consumption of resources and larger worst-case completion time.[0006]A problem of keeping multiple caches consistent with...

AI Technical Summary

Benefits of technology

This approach enhances system performance by dynamically adjusting consistency levels, reducing resource consumption and latency, and improving scalability by tailoring consistency management to the changing needs of objects throughout their lifetime.

Problems solved by technology

How the updating of the different copies is coordinated leads to different levels of consistency among the copies, in return for different costs to perform that coordination.

Typically, a stronger consistency, with closer coordination between peer cache updates, results in a larger consumption of resources and larger worst-case completion time.

A problem of keeping multiple caches consistent with each other is evident in processor caches for multiprocessors and file caches for distributed file systems.

Techniques that work well given those facilities are simply not practical for distributed applications such as Web caches.

This weak consistency scheme, with the clients checking with the server (see e.g., J. Howard, M. Kazar, S. Menees, D. Nichols, M. Satyanarayanan, R. Sidebotham, and M. West in “Scale and performance in a distributed file system”, ACM Transactions on Computer Systems, 6(1):51-81, February 1988), can have significant overhead.

One problem with implementing applications in prior art consistency schemes emerges from the fact that these schemes are not sensitive to the variations in the consistency needs of an object during its lifetime.

The consistency mechanisms are not flexible enough to adapt to the consistency requirements of the object based on its state.

Maintaining objects at strong consistency levels is costly in terms of the load on both the consistency coordinator and the caches.

This heavy load results in the end user experiencing high latencies.

Further, if the load of the consistency management is bursty (as it often is), it will suddenly increase the load on the system.

This would not only increase the average latency to unacceptably high values, but might also result in problems such as buffer overflows causing requests to be rejected by the system.

Therefore this approach of maintaining strong consistency for all objects does not scale well with the increasing number of objects.

Images