In the late 1990s at the height of the “dot-com era”, investors of all sorts (including many who should have known much, much better) were throwing money at any Internet-based opportunity they could find. A significant amount of that funding went to a category of Internet players that came to be known collectively as storage service providers (SSPs).
This group came into the marketplace, collected investors’ dollars, pounds, francs, and yen, stayed for a brief while and then, almost as suddenly as they appeared, began to fall by the wayside. By early 2001, these dot-coms evaporated into “dot-gones” and the investment portfolios of the venture capitalists and other investors who had backed them were left hemorrhaging red ink.
For the most part these SSPs died on the vine for the most legitimate of reasons—they were not providing a service the marketplace thought was particularly valuable. In many cases, what they offered was just added mass storage, located off-site and accessed either via the Internet or in extreme cases, via a leased line. In other words, these were essentially co-location sites offering little in the way of value-add services.
Co-los of course can be quite useful, particularly if they are set up as part of a disaster recovery (DR) strategy or increasingly the case nowadays when the primary data center has been maxed out in terms of available power or floor space.
Back in the Day
In the early part of this decade, however, while there certainly were some remote DR sites, few if any data centers were strapped when it came to power or space requirements. In any case, most of the dot-com storage providers offered nothing like that sort of value. They were simply businesses whose principal competency lay in an ability to build, lease or buy buildings efficiently, and then to provide those sites with adequate power and security.
After that, so the theory went, with enough customers they could buy storage systems by the boxcar and thus drive down their hardware prices. Their profit margins were covered, to a large degree, by the extent to which they could commoditize their storage purchases.
While most SSPs turned out to be little more than off-site storage with some limited storage management thrown in, there were of course some interesting exceptions. Storage networks and storability come easily to mind. Storage networks provided not only storage but sophisticated storage management as well. They cherry-picked the best from a number of vendors, threw their own management software on top of that, and offered a complete package well before any such coherent solution was available from most storage vendors.
Storability, as I recall, had a model that was the inverse of what everybody else was doing. They offered a Network operations center (NOC) that reached out and remotely managed their customers’ data centers. Storage networks ultimately crashed and burned, badly singeing all its investors. Storability was eventually purchased for pennies on the dollar by StorageTek, which of course was in turn acquired by Sun two years ago.
It’s pretty clear that most of these early attempts at off-site storage were doomed from the start, but it’s important to understand that the reasons for their failure often had nothing to do with either the technology or with the business value of the service they were providing. It was a combination of “data center sociology” and a turn-of-the-century sense of corporate governance that were the SSPs’ downfall.
First, the mindset of most IT managers had not advanced to the point where they could tolerate housing and servicing data off site. For them the SSP was more of a threat to their existing bailiwick than it was a help. For them, the SSP was a form of outsourcing; one that would potentially deploy jobs and budgets away from the data centers and out beyond the range of their control.