Cloud Zone is brought to you in partnership with:

Clustrix has developed ClustrixDB, the industry’s first scale-out SQL database engineered for the cloud. ClustrixDB is uniquely and ideally suited to handle massive volumes of ACID-compliant transactional workloads. Lisa is a DZone MVB and is not an employee of DZone and has posted 13 posts at DZone. You can read more from them at their website. View Full User Profile

DBaaS: Solving the Cloud Business Continuity Conundrum

07.01.2014
| 4371 views |
  • submit to reddit
Today’s enterprises—especially those in the financial and e-commerce industries—are expected to be “always-on”; you never know what time of day or night your customers are going to make a transaction. At the same time, these companies are increasingly turning to the cloud to improve efficiency, scalability, and reduce costs while dealing with the tremendous and varying amounts of data they need to run their business.

Enter database as a service (DBaaS), which helps businesses deploy new databases in the cloud quickly, cost-effectively, and without the hassles of database administration. MarketsandMarkets forecasts the cloud database and DBaaS market, including software and services, to grow from $1.07 billion in 2014 to $14.05 billion by 2019, at a CAGR of 67.3 percent.

But what happens when there’s an issue such as a natural disaster or even a catastrophic power failure? While all business continuity plans include disaster recovery, they usually don’t accurately take into account how long a business might be down while databases are brought back online. Traditional MySQL databases can take a long time to recover, as data is restored to the primary database server, or a replication slave is promoted to master and the database is restarted.

If there is a failure in the cloud, the database must be able to recover quickly. Databases based on NewSQL provide simpler and far more robust high-availability approaches that speed recovery. For example, ClustrixDB is built with automated fault tolerance and recovery. It also has self-healing capabilities, which re-protect the cluster after failure recovery to be able to survive a successive failure prior to replacing failed components from the initial failure.

With its self-managing architecture, ClustrixDB is built with redundancy within a cluster and dynamically distributes the database across multiple nodes for maximum reliability. ClustrixDB maintains multiple copies of all data within a cluster so if a node fails, the lost copies are regenerated and redistributed across the cluster. The cluster survives and automatically recovers from the failure in seconds (not hours) and keeps running. ClustrixDB is also built with fast parallel backup and restore for disaster recovery, with all servers restored concurrently, instead of one by one as in a traditional environment. This feature takes a consistent snapshot of the database, and the backup time does not increase with size of the database.

Learn more about how ClustrixDB can offer five9s reliability and high fault tolerance with the flexibility and economics of the cloud to effectively solve the cloud business continuity conundrum.

 

Published at DZone with permission of Lisa Schultz, author and DZone MVB. (source)

(Note: Opinions expressed in this article and its replies are the opinions of their respective authors and not those of DZone, Inc.)