Amazon Improves Reliability for Its Cloud-based Database
Amazon Web Services has added a new high-availability feature called Multi-Availability Zone (Multi-AZ) to its cloud-based Relational Database Service (RDS), the company said on Tuesday.
The addition of Multi-AZ provides better database availability by automatically configuring a standby copy of the database, which is stored in a different physical location from the original in Amazon's cloud. Updates to the primary database are automatically replicated to the copy, according to Amazon. A drawback is that users may experience more latency compared to a database that doesn't use Multi-AZ, Amazon said in an FAQ..
The copy takes over during an outage caused by, for example, a network or a storage failure or during planned maintenance. From start to finish, the failover between the two databases typically completes within five minutes, according to Amazon.
Adding Multi-AZ doubles the hourly costs users pay for each database instance to between US$0.22 and $6.82, depending on the location and the performance metrics of the instance. Users also have to pay for primary and backup storage, the number of requests, and data transfer in and out of Amazon's cloud.
Amazon RDS is based on MySQL 5.1 and is still in beta.
Amazon is also working on a second feature that will replicate data to improve scalability by using multiple database instances at the same time. So-called Read Replicas will make it easier to take advantage of MySQL's native, asynchronous replication to dynamically add instances to handle large amounts of incoming read traffic, according to Amazon, which didn't provide when the feature would become available.