Joe Tucci, EMC chairman and CEO, in his EMC World keynote address last week said that storage is a primary driver and virtualization the key technology for cloud adoption and that the complexity of storing large amounts of data will require customers to adopt the cloud to make that data readily accessible.
Tucci was talking primarily about customer adoption of private clouds. EMC’s latest storage offering is VPLEX, which is the company’s new technology for pushing storage virtualization towards private clouds. The technology enables federation of distributed storage over long distance thus allowing for seamless access and migration of data between geographically separated data centers. Tucci commented:
“So at the end of the day, you’ll have thousands of private clouds, and hundreds of public clouds”
We validated this statement in our own research; it will be a hybrid cloud storage world for many years to come with a lot of that storage living in private clouds.
Many large enterprises deal with petabytes of data and their processing. The CIO of one Fortune 500 financial company told us that the company currently has 12 petabytes of data, growing 40 percent annually. Even if you could host some of this information and process it on an external cloud, the result sets that come out of these applications are often in terabytes. The more access you need to your data, the more you get anchored to a particular location particularly if low latency is an important consideration.
Not all data access patterns are well suited to the cloud, particularly if there are large distances to cover. In such cases, bandwidth becomes not only a challenge but a financial consideration. The CIO joked: “The fastest way to send a terabyte of data from London to New York is a jumbo jet.” The EMC VPLEX products are aimed at solving this problem. But issues such as disk drive reliability, bandwidth limitations, and speed of light are natural constraints. These may ultimately prove to be the biggest obstacles to developing cloud distributed storage architectures that can be as reliable as and more cost-effective than storage area networks within a data center. That said, what we do need is cloud storage architectures such as VPLEX which can optimize the use of distributed data centers so that the overall efficiency and utilization goes up dramatically thereby reducing costs.
Cloud storage isn’t about to replace the storage network in the data center any time soon, at least not for data-intensive, high-performance, low-response time, transactional applications and for mission-critical data. But we will see many use cases where companies and organizations of all sizes will augment their on-premise storage with cloud storage (potentially from various vendors) in a hybrid model deployment. However, hybrid models tend to bring interoperability issues and the need to deal with different tools, API’s, management infrastructure etc. Customers would like to use their familiar existing technologies, tools, and user interfaces to handle hybrid cloud scenarios seamlessly and securely. StorSimple is offering innovative hybrid storage solutions that makes cloud storage (from multiple vendors such as EMC, IronMoutain, Microsoft, and Amazon) appear like local data center storage that integrates into customers’ existing storage and data management tools.
The popular storage use cases tend to be infrequently accessed data scenarios including archiving, backup, DR, and offsite data protection.
The question however to be considered is whether cloud archiving will be cheaper and more reliable over a long period of time (typically 7 years for archiving needs) than the old-fashioned tape archiving solutions, particularly for large volumes of data.
Many cloud storage companies charge for the amount of data transferred in and out of their service (even as some vendors offer a fixed fee for unlimited storage) so costs can add up quickly. OpSource’s most recent storage offering charges only for bandwidth when files are retrieved, not when files are uploaded to their cloud storage, potentially saving half the costs. Some offering full end-to-end encryption of data, both in motion and at rest. Others don’t. There are plenty of enterprise cloud storage options today (Amazon, RackSpace, Iron Mountain, ATT, Microsoft, Google etc), so it pays to shop around for the best combination of features, cost, performance, reliability, and security that meet your business needs.
Time taken to archive/backup can potentially become an issue. EMC-owned Mozy is introducing in its next release the ability for users to backup their data locally in addition to cloud-based backup.
A more important question is if and how you can get your data back if the cloud vendor goes out of business—a not very unlikely scenario in a nascent market that is sure to experience some industry shakeout in the next few years.
Read More EMC tempts storage buyers into its federation
Cloud Storage will be limited by drive reliability and bandwidth
Can Cloud Computing Replace Your Storage Network?
Disk and Cloud emerge as archiving solutions
CLOUD EVENT NEXT WEEK: We are co-hosting with Ness Technologies a Software Executive Networking Dinner, “Achieving Technology Leadership in the Cloud,” Wednesday May 26 at 6:00pm at Hiller Aviation Museum in San Carlos, CA. If you are based in the Bay Area or are visiting, please join us for the dinner and enhance the quality of the interactions and networking. Please drop me a note if you can make it to the dinner.
Kamesh Pemmaraju heads cloud research at Sand Hill Group. He welcomes your comments, opinions, and questions. Drop in a line to kamesh@sandhill.com.