Experts have almost unanimously agreed that big data would be too costly to move to the cloud. But suppose remote backup copies could be deployed to the cloud, what would that mean?
Superficially, the public cloud seems like the perfect place to implement big data analytics, DevOps and related actions. The hyper scalability that cloud infrastructure provides makes it ideal for batch jobs especially pay for powerful infrastructure need to process the data, and then shift back to your normal package when youre finished.
Little surprise that AWS moved to ElasticMapReduce and Redshift data warehousing once it was through with S3 and EC2 implementation. However, theres just one problem: moving big data is a real pain. From the prohibitive cost of acquiring enough bandwidth to the unfulfilled need for real-time analytics, there is still some work to do in this area.
Cloud data and on-premise data has still not fully synchronized, particularly as regards big data, which remains a problem because enterprises are now moving closer and closer to real-time analytics. While this wont change for a while, what if there was a way to circumvent it?
Virtualization and Copy Data Management
Many enterprises have implemented backup and disaster recovery solutions in the cloud. Rather than have this data sitting idly in the cloud, waiting to kick in when there has been a disaster, why cant virtual copies of that data drive big data tests, development or analytics from the cloud?
Until now, cloud DR and backup was a solution for small to medium businesses alone larger businesses driven by high availability relied on deployment to special backup data centers in other site. Data would replicate to these sites periodically, but at a high cost. However, this too only kicks in should there be a disaster.
Case study
The best example of the above proposition can be evident with Actifios idea of using one, constantly updated enterprise data copy to create virtual copies used for backup, DR, analytics and DevOps and testing within the cloud environment.
The company, founded in 2009, received $ 100 million in March 2014 from key stakeholders to create and avail a platform through which a single copy (christened the golden copy) of an enterprises data can be virtualized and used for multiple purposes.
Virtual data management would address a key pain point with enterprise data requirements. Enterprise data is growing exponentially, with Hadoop analytics, data warehousing and accelerated app dev and testing placing an even bigger burden in enterprises storage infrastructure.
The importance of cloud backup and DR cannot be over-emphasized, and even organizations with big data would be well served to look into its potential. From hyper-scalability and cost efficiency to remote access and high availability, the capabilities of online backup and DR are without number.
Once you have a single copy to DR and backup, it makes sense to create virtual copies for the remaining applications, rather than making multiple physical copies for analytics, development and testing, which creates a need for even more storage space in the cloud or on-premise.
Conclusion
Given the potential of data virtualization, its only a matter of time before Microsoft, Google and AWS, the public cloud giants, venture into virtual management. Cloud copy management has huge potential, given that big data analytics for development and testing is one of the main uses of cloud infrastructure as far as big data is concerned.