Consider the (Out) Source

As a young networking engineer, I was always infatuated withthe concept of bandwidth on demand. As my customers consumed more bandwidth, Icould add incremental units of networking capacity to my network. Rather thanhaving to lease entire new circuits, I could ask that the capacity of myexisting lines be upgraded to meet my needs. It sure sounded good.

For all the years that I managed an enterprise network,bandwidth-on-demand was long on promise but short on reality. Whether it wasframe relay, fractional T1, or ISDN, I never found a economical means toprovision bandwidth-on-demand without paying-through-the-nose.

So, what to make of recent moves to providestorage-on-demand?

First, let’s consider the idea. The market seems to divideinto those who provide storage remotely, those who combine remote andenterprise storage, and those who build storage infrastructure for a companyand then manage it remotely. There’s activity in all these areas and, while itseems that mostly large companies are succeeding with remote storage, manyvendors are touting storage as a service. Instead of racks and rooms of disks,some companies appear to be taking advantage of outsourced storage.

Consider IBM’s recent announcement that it will offer datastorage on a pay-per-use plan. IBM’s announcement is just one part of anoverall application-service-provider strategy where technology is delivered aspackaged services. The idea is to give enterprises access to fundamentalinfrastructure without those companies having to worry about new technologies,management, or up-to-the-minute developments in storage. Electronic commerce,extranet, and other e-business applications are also creating an extraordinaryneed for fast, secure access to stored data. With storage management costsgrowing and a shortage of trained staff to deploy new storage technologies,many companies are looking to network-delivered storage solutions.

IBM’s idea is to make storage capacity and management acommodity -- allowing customers to get what they need only as they need it.According to IBM, companies will be able to choose to have IBM manage theservices remotely at the customer premise or at one of its 175 data centers.

IBM is not the first to let companies outsource storageneeds. For some time, StorageNetworks rented access to its own storage systemsat more than 20 data centers in several sites around the country.

Another approach has been to build deploy-on-demand storage.Some vendors -- such as Sun and Compaq -- have touted the availability ofstorage capacity programs as "utilities," but to my eyes these havemostly seemed like typical storage leasing programs dressed up in fancyclothes.

IBM and StorageNetworks’ approach may work well fore-commerce companies, but I’d wager that many IT managers in largeorganizations are wary about security. Can an IT manager convince a CEO topermit the storage of crucial corporate data at a remote site where it could besharing the same RAID array as a competitor? Considering the security andavailability issues that fully outsourced data raises, it’s understandable thatsome companies have looked at different strategies for outsourcing.

One intriguing approach is that of a new company calledStorability. This new competitor attempts to provide a start-to-finish approachto outsourced storage. First, Storability works with an enterprise to evaluatea company's data storage needs, then it recommends a system that addressesthose needs, and then moves from the existing environment to the new storageinfrastructure. Storability proposes both on-site and dedicated off-sitesolutions and offers a service to manage the new infrastructure remotely. It’sthis combination of customer involvement, local or dedicated solutions, andservice-level agreements that make Storability an attractive alternative tosimple outsourcing.

It’s fascinating to consider what the so-called experts haveto say about the marketplace. Almost everyone agrees that in 1999 the storageservices marketplace was a miniscule $10 million. Still, projections for thefuture are almost impossibly rosy. For instance, Dataquest suggests that themarket will grow to $5 billion by 2003! Not to be outdone, Strategic Researchthinks the market will blossom to $7.8 billion in the same period. IDC seeseven more potential, predicting the market will be as large as $8 billion in2003.

Are theexperts right? I’m skeptical. It’s clear that enterprise storage is becoming abigger problem for all large enterprises, but I’m not sure that there’senormous demand to put critical data into the hands of a third party. So far,storage-on-demand seems to be a concept heavy on promise but -- likebandwidth-on-demand -- short on real products. While solutions like Storabilitymay meet many managers’ concerns, I still doubt that the market will explode atthe pace industry experts predict. --MarkMcFadden is a consultant and is communications director for the CommercialInternet eXchange (Washington). Contact him at mcfadden@cix.org.