Microsoft has provided erasure coding in Windows Azure Storage. Many experts have proved that the erasure coding has many benefits compared with the replicated. What are your thoughts?
If you're asking whether you think cloud storage based on erasure coding techniques (e.g. Local Reconstruction Codes) will be used more often in the future over over replication / reconstruction techniques for high performance / availability, I think there will continue to be improvement and innovation in the area and the choice will come down to what will provide the best products for the providers customer base.
In terms of what customers want from cloud storage in the future - they probably don't care about the algorithms, they care about the cost / durability / capacity / performance.
If you're asking whether you think cloud storage based on erasure coding techniques (e.g. Local Reconstruction Codes) will be used more often in the future over over replication / reconstruction techniques for high performance / availability, I think there will continue to be improvement and innovation in the area and the choice will come down to what will provide the best products for the providers customer base.
In terms of what customers want from cloud storage in the future - they probably don't care about the algorithms, they care about the cost / durability / capacity / performance.
I think if you look at the history of NAS storage, including the evolutionary and revolutionary changes that have occurred there, you can find a pattern that will mostly repeat itself in the cloud. Storage will move from WAN to Cloud, just as it did from LAN to WAN as the reliability, performance and security concerns are addressed by technological innovations. The timing is different, and the problems are different, but overall history will repeat, and repeat again.
Erasure coding is one of those technical solutions that addresses a reliability concern in a new environment. If I wanted to find some way to write a paper on this, I think I would try to draw parallels with changes that have occurred in storage in the last decade, then research the open issues in the new environment that stand in the way. That should provide a roadmap for open research problems in cloud storage, and perhaps some insights on how it might evolve.
Actually, I think the cost of using cloud storage should be lower than building and maintaining your own storage, assuming you need decent amount of reliability and availability.
How many people are running their own web servers for their web sites? The same will be happening with storage.
The future of cloud storage is brighter! As individuals and companies all over the world continue to create and own a straggling amount of data each year, one can only conclude that cloud storage adoption will be more prevalent. Security issues related to cloud storage predictably will be overweighted by its advantages. Innovation shall also provide solutions currently hindering its full adoption, in terms of security, access, and energy consumption related to current cloud storage, as highlighted in Corcoran (2012).
I expect that cloud storage as an issue for a customer moves to a backend layer which will no more be recognized by the enduser. Main reason for this is the strong trend of the move of applications (all types) into the cloud. In future we will book a service and storage will be a part of it. The technology how data is stored will play a really minor role. Apple (consumer) or salesforce.com (business) are here good examples