Choosing Database Architecture for the Cloud

Technologies have advanced so much that the software architect doesn’t have to worry about the finer details any longer. The limitless infrastructure and the ever-presence of applications are making solutions feasible that were not even dreamed about a few years ago.

The disruption of the traditional database gave us deliverance from the monoliths, and blew open the architectures and philosophies to consider before deciding where to put data.

DBTA recently held a webinar featuring Seeta Somagani, solutions architect at VoltDB, who explored critical challenges to consider when choosing the database and architecture for deployment in the cloud.

The programmer’s job is to squeeze as much computing out as possible to make dreams a reality, Somagani explained.

Arguments to tackle complexity:

  • Program Length !~ effort to design
  • Avoid clever tricks
  • Solutions must be nicely factored

The cloud architect has complex issues of their own to figure out such as orders of scale, Overwhelming choice of technology, and ever-higher expectations of availability.

Databases for the cloud need to be scalable, be a cloud deployment model, and have service-oriented architecture.

There are three database choices, according to Somagani. This includes monolith, distributed, and deconstructed.

VoltDB can offer the following:

  • ACID Consistency-
  • Jepsen Testing
  • High Performance
  • Transactions
  • Serializable transaction model
  • Partition data & processing
  • Advanced Materialized Views
  • JSON Columns
  • XDCR
  • Stream + Batch processing
  • Horizontally Scalable
  • Continuous queries Relational + Semi-structured data
  • Cloud-ready

An archived on-demand replay of this webinar is available here.