The fifth annual cloud conference and awards December 11, 2014 - San Francisco
cloud computing for healthcare

Scaling MySQL in the Cloud

SESSION DETAILS

Date: December 12th 2012
Time: 05:20pm PST
Location: Willow Glen Room


SESSION ABSTRACT

MySQL doesn’t automatically scale in the cloud. In fact, the less powerful machines that are available force you to use partitioning and sharding techniques earlier. Achieving High Availability on the cloud with large data sets is a challenge as cloud-hosted servers are less reliable and predictable than dedicated hardware. However, it possible to deploy MySQL using a shared disk approach. This architecture is based on a cluster of MySQL servers and multiple storage nodes such that multiple MySQL servers can update and query the same physical database. Scaling is achieved by adding MySQL servers and storage nodes to the cluster (without the need to partition or shard the data). This approach provides High Availability as if a MySQL server fails; the applications continue to operate with the surviving servers without downtime. As the data is mirrored on multiple storage nodes, if a storage node breaks, the cluster continues to operate. This session will demonstrate and explain the different methods to scale MySQL in the cloud as well as the technology, the architecture and the use cases of ScaleDB that provides the infrastructure for a dynamic cluster of MySQL and storage instances.

SPEAKER

Moshe Shadmon , Founder, CTO , ScaleDB Inc.
Moshe Shadmon is the founder and CTO of ScaleDB. Moshe has more than 20 years of experience in both technical and management roles. Moshe’s technical experience has been in the areas of database clustering, high performance indexing solutions as well as the development of a unified solution for structured and semi-structured data. Since 2006, Moshe has become increasingly interested in the opportunity presented by the open source databases and founded ScaleDB to develop a database engine for MySQL targeting performance, efficiency and scalability with the largest data sets.
Fifth Annual Event