CoLo Cage

The Future For CoLo

CoLo CageTimes are looking good for co-location (co-lo) facilities.  As commercial entities realize that building yet another data center for their own use is not exactly cost-effective, they look towards the value of using an external one for some or all of their needs.

For some, the use of the “XaaS” (whatever-it-may-be as a service) model will be the way that they will go.  However, there will always be the “server huggers”: those who realize that owning the facility is now counter-productive, but who still want to own the hardware and software stack within the facility.  For these people, letting someone else worry about power smoothing and distribution, environmental monitoring, auxiliary power generation, facility cooling, and connectivity makes a lot of sense.

And, then of course, there are those who want to be in the XaaS market — but as a provider, not as a user.

Providing services to the general public and commercial organizations can be fraught with danger — particularly if you are new to market with a relatively new idea. Will everyone go for The Next Great Idea? If so, going it alone and building a full data center requires a crystal ball of the largest, clearest magnificence:  get it wrong and too many customers appear, and you find yourself needing to splash out on another facility far too early.  Get it wrong, and not enough customers come to you, and you find yourself trying to keep a vast empty data center going with low cash flows.

No — better to go for co-lo, and start at the size you need with the knowledge that you can grow (or shrink) as your needs change.  Funding the relatively small amount of money needed up front to get off the ground can then make it so that cash flows come through the hockey-stick curve rapidly, and the positive, profitable cash flow can then be used to fund the incremental increases in space that the service provider needs as customer volumes increase.

So far, hopefully all pretty obvious.

Co-lo does, however, offer both users and service providers further advantages.  As there will be many customers within a single facility, they are all capable of interacting at data center speed.  Therefore, a service provider housed in the same co-lo facility as their own customer can pretty much forget about any discussions on data latency — core network speeds will mean that this will be down in the microseconds for a well-architected platform, rather than the milliseconds.

Even where the service provider is in a different physical facility than the customer, but with the same co-lo provider, the interconnectivity between the facilities will minimize latency far below what an organization could hope to get through the use of their own WAN connectivity.  Even between co-lo data centers owned by different companies, the amount and quality of connectivity in place between the two facilities will still outperform the vast majority of in-house data centers.  Combine all of this with judicious use of WAN acceleration from the co-lo facility to the end user in the headquarters and/or remote offices of the end customer and a pretty well performing system should be possible.

For co-lo providers like Interxion and Equinix, this points to a need to not just be a facility; nor even a fully managed, intelligent facility.  What it needs is what Interxion calls a “community of customers, partners and suppliers”.  By working as a neutral party between all three, it can help advise all its customers on how best to approach putting together a suitable platform.  They can also help with advising on which of its partners can provide services that could make life a lot easier for a customer who may — through no fault of their own, apart from lack of time to check into everything available out there — be intent on re-inventing the wheel. In some cases, this may also be advising on what services are provided from outside its facilities — such as the use of Google or Bing Maps in mash-ups, rather than buying in mapping capabilities

This breeds a new type of professional service: the co-lo provider which does not provide technology services per se over and beyond the facility itself, nor provides systems integration services, but does provide the skills to mix and match the right customers with the right service providers.

Backing this up with higher-level advice on how to approach architecting the end-customer’s own equipment to make the most of the data center interconnect speed and minimize latencies to provide the best possible end-user experience should be of real value to all co-lo customers – particularly commercial companies struggling to fully understand the complexities of next  generation platforms.

If you are looking for co-lo, Quocirca heavily recommends that you ask prospective providers whether they plan to offer such a neutral brokerage service – and if not, walk away.

Image credit: ChrisDag (flickr) – CC-BY-2.0