|FoundationDB Adds Multi-region Support|
|Written by Kay Ewbank|
|Monday, 26 November 2018|
A new version of FoundationDB has been released with support for multi-regions within single clusters, faster failover recovery and improved TLS.
This is the first major release since FoundationDB was made open source. FoundationDB was bought by Apple in 2015, and the core of the software was open sourced in April 2018.
FoundationDB is very fast, supports ACID-compliant transactions even though it’s NoSQL, and scales well. It's a distributed database designed to run on clusters of commodity servers to handle large volumes of structured data. Data is organized as an ordered key-value store, and you interact with it using API language binding.
The changes to the current version start with the native multi-region support. This aims to make databases better available globally. Seamless failover between regions is now possible, so that if an entire region goes down, your users should still not experience a service interruption. You can set up FoundationDB so that these features are used but clients still have low-latency, single-region writes.
If you choose a multi-region configuration for failovers, one region will operate as the write authority by default. If an availability zone has a problem, FoundationDB automatically transfers authority to the other region without any data loss. If you want to run FoundationDB in a single region, a cluster can now be configured to intelligently make use of multiple availability zones in case of problems.
In order for the multi-region support to work, FoundationDB now recognizes region locality. When data is written to a disk, whether on a storage server or transaction log, information is included on the region the disk is in. This makes various things possible, including being able to choose operations that are local to the region by preference.
The regional awareness is also used to conserve bandwidth when replicating data over expensive, high-latency intra-region links. FoundationDB sends a single copy of data between regions and then redistributes the data to replicas on the remote side.
Regional failure modes have also been added. In the past, if a machine failed, the assumption was that the failure was permanent. and data was immediately copied onto other machines. This was problematic in the case of a regional or availability zone failure, because such problems are more frequently temporary, and expensive to recover from, since copying the entire database's contents over the WAN is expensive. In the case of a failure across a region or availability zone boundary, FoundationDB can now be configured to treat the problem as temporary and keep local copies until the failed region comes back online.
or email your comment to: firstname.lastname@example.org
|Last Updated ( Monday, 26 November 2018 )|