With its Centera storage system, EMC was the standard-bearer for content-addressable storage; but the proprietary...
nature of the product is wearing thin.
Around 2003, we seemed to be on the threshold of a new storage era that had the potential to revolutionize how we stored data and could use it more effectively in our businesses. The optimism was encouraged by a new concept called content-addressable storage (CAS) and EMC was its key proponent with its Centera system. Centera was an object-based storage system that provided extended metadata attributes to give users more control over and insight into data.
EMC successfully recruited independent software vendors (ISVs) to support Centera’s application programming interfaces (APIs), which required more work but provided deeper integration between the applications and hardware.
It was a powerful vision: a storage system that removed the complexities and many of the limitations of storage-area network (SAN) and network-attached storage (NAS). You could scale Centera to any size without increasing the amount of administration. You could leverage the extended metadata to enforce policies on content. The most prominent example of this is its write-once read-many (WORM) capability that prevents stored documents from being altered so they can be used as evidence in litigation or for audits.
EMC took a very smart and practical approach to the Centera go-to-market, knocking on the doors of those industries that had relied heavily on optical disks. The ROI calculations suggested that even though there would be an increase in infrastructure spending, the operational cost reductions could save users hundreds of thousands or even millions of dollars over time. And with the backing of the ISVs that supported Centera, it was relatively easy to sell hardware/software solutions. Those efforts were bolstered by EMC’s great reputation and excitement around the new storage platform that seemed to present a great vision for the future of storage.
And then the government passed a number of regulations, including SEC Rule 17a-4, Sarbanes-Oxley, the FDA’s 21 CFR Part 11 and new HIPAA rules, propelling Centera into a billion dollar business. The timing was perfect.
Other storage vendors and startups joined the fray. NetApp added WORM functionality and Hewlett-Packard, IBM and Sun all made efforts but ultimately failed in this space. Hitachi acquired a startup called Archivas but missed the market. Other startups, including Caringo and Permabit, found this was a big vendor’s game. In fact, it was only EMC’s game and everyone else either did OK (NetApp) or terribly (pretty much everyone else).
But trouble began to brew in paradise. Some people questioned the viability of Centera, while others said it was slow and expensive, proprietary, corrupted data and wasn’t that easy to manage.
Over time, Centera went from being the hottest (and coolest) storage system in the market (and often one of its most controversial), to almost being just an afterthought. It became yesterday’s news.
Today there appears to be little to no continued engineering for Centera, even though there’s an estimated 600 petabytes of it in the market. One Centera user told me he believes the latest release of it will never come out. EMC is going to replace this particular user’s Centera with Atmos, and provide the same functionality with the exception of WORM and immutability, which this user’s environment doesn’t require. However, depending on when or if Atmos ever supports WORM, the customers using the Compliance Editions of Centera may be stuck.
Some of the objections to Centera may have been right. However, the main problem is Centera’s proprietary APIs. Both EMC and ISVs locked-in customers, so they’ve avoided price erosion or competitive threats. Some users told me that unlike other storage systems, Centera has actually gone up in price.
It’s tough to migrate data from a Centera to a competitive platform. I spoke to another user whose firm took 18 months to migrate off Centera, an effort that included custom development and pressuring its ISV to support the move.
There should be some outrage about this situation. Centera is a proprietary system, which should be unacceptable in today’s data center. It’s true that Centera users made their choices voluntarily, but it now looks like they’re locked into a platform that has no long-term future.
EMC should offer a way to make Centera non-proprietary and provide free migration services to a more sustainable and open platform. Meanwhile, the storage vendors that missed the boat when Centera was on the rise (all of them) will get a second chance as Centera’s star fades.
BIO: Tony Asaro is senior analyst and founder of Voices of IT.
- Demystifying storage performance metrics –ComputerWeekly.com
- Containers and storage 101: The fundamentals of container storage –ComputerWeekly.com
- All-Flash: The Essential Guide –ComputerWeekly.com
- Hybrid Flash: The Essential guide –ComputerWeekly.com