Back at the peak of the Internet boom, Venture Capital firms threw huge amounts of a money in a black hole called storage service providers (SSP). At the time, many assumed that the SSP model would become the normal way for enterprises to manage storage. But along came the crash, pouring a bucket of cold realism over the whole SSP industry, and suddenly sanity prevailed! Overnight storage service providers were rather unkindly dubbed - suddenly software providers. The hundreds of millions of dollars poured into the industry by VCs and investors, seemed to melt away along with the companies that wasted it.
Hindsight exposed two critical problems with the whole concept of SSPs:
- Strangely enough, most companies didn't particularly like the idea of giving up their storage because it meant losing control of their data (as my famous relative Homer would say, Doh!)
- The infrastructure required to move critical storage offsite didn't exist, and the cost of building the infrastructure was prohibitive.
So fast forward to 2007 and amazingly, the idea of the SSP seems to be rising from the grave, it's looking a bit shabby, but not too bad for a moldering corpse. At least somebody had the decency to dress the corpse in a nice new "Storage As A Service" (SAAS) T-shirt. The question is, has anything changed to make outsourced storage a success now? First, let's look at the two major problems that killed the SSP:
- In an age were corporate data is a key company asset and paranoia about data leaks is rampant it seems unlikely that many large companies are going to be any more sanguine about outsourcing their data in 2007, than they were in 2000.
- As far as the infrastructure is concerned nothing has changed, if anything it's got worse, with widespread use of content like video downloads in danger of bringing the public Internet to it's knees.
Undoubtedly some companies will embrace the idea of SAAS, but only as an internal discipline to improve the management of storage. I think the chances of large companies outsourcing critical storage assets is somewhere between slim and none.
The more interesting question is whether other less sensitive and mission-critical storage applications might get outsourced. One area of interest is archiving of data, I've talked to several companies in recent weeks who are considering getting into this market with outsourced offerings. Archiving might just make sense because:
- Archiving is not mission critical, and the privacy aspects could be addressed with strong encryption of the data before it is sent offsite.
- Archiving is not performance sensitive, so long as there is enough bandwidth to move the data then there's no problem.
- Archiving capabilities might be built into existing and relatively popular offsite backup systems which already have the data.
- Many companies already send data offsite in the form of tape, so the idea is not completely foreign to them.
The problem with outsourced archiving is that moving the archive offsite only addresses a small part of the problem. Today, companies can find a variety of hardware products from vendors like EMC, HDS, IBM, HP, Caringo etc. that can provide archive ready storage solutions. The problem is not "where to store archive data", the problem is "what data to archive" and nobody has a good a solution to that yet.
I also think there is a considerable opportunity for providing outsourced storage and services to home users. But if I were a VC listening to a startup that thinks the SSP model will work any better in 2007 than it did in 2000, I'd be thinking of George Santayana and his famous remark "Those who cannot remember the past are condemned to repeat it."
Posted by: Nik Simpson