Sign In

Communications of the ACM

Research highlights

Technical Perspective: Getting Consensus For Data Replication

Data in a cloud computing system should be highly available. That is, whenever you connect to the system, the data you stored there should be ready to use. The standard mechanism to accomplish this—data replication—involves maintaining multiple copies of each user's data.

By itself, replicating data does not solve the problem, because the copy of data that is available might be stale. To see why, consider a system that maintains three copies of file x: x1, x2, and x3. Suppose x1 and x2 are currently available and x3 is down. If a program updates x, the system writes the update to x1 and x2, but not to x3 because it is down. No problem, since the system still has two correct copies. Next, suppose x1 and x2 fail, and then x3 recovers. Now there is a problem. If a user reads x, the best the system can do is return the value of x3. But that copy is stale—it does not have the latest update.


No entries found

Log in to Read the Full Article

Sign In

Sign in using your ACM Web Account username and password to access premium content if you are an ACM member, Communications subscriber or Digital Library subscriber.

Need Access?

Please select one of the options below for access to premium content and features.

Create a Web Account

If you are already an ACM member, Communications subscriber, or Digital Library subscriber, please set up a web account to access premium content on this site.

Join the ACM

Become a member to take full advantage of ACM's outstanding computing information resources, networking opportunities, and other benefits.

Subscribe to Communications of the ACM Magazine

Get full access to 50+ years of CACM content and receive the print version of the magazine monthly.

Purchase the Article

Non-members can purchase this article or a copy of the magazine in which it appears.
Sign In for Full Access
» Forgot Password? » Create an ACM Web Account