<img src="http://www.gblwebcen.com/41894.png" style="display:none;">

Volta Data Centres Blog

Why would anyone want to own their own infrastructure? - by Phil Alsop, Editor, DCS Europe Published

Posted by Volta Newsroom on 02-May-2017 09:29:21

You have a great idea for starting a new business. As part of the launch process, you put together the necessary business plan. You include the price of purchasing a small power station, a small telephone exchange, and both a water treatment and sewage treatment plant…Of course you don’t!

it_infrastructure-5.jpg

In reality, you are more than happy to leave the electricity, water and telecoms provision to the experts, who will carry out all the necessary connections, maintain their respective utility services, sort out any problems, even upgrade/improve the infrastructure over time and all for a very affordable annual fee, which they’ll even let you pay in monthly instalments.

Similarly, if you are providing your employees with company cars, you could go out there and purchase some vehicles at a considerable capital cost, and then have to manage the fleet, paying for all the repairs and, eventually, having to purchase replacement vehicles. Or, like virtually all businesses, you could just lease the necessary vehicles, for a fixed monthly fee, including all maintenance, and have them replaced every three years.

When it comes to the data centre and IT infrastructure, why is it then that so many organisations still insist on purchasing, operating and maintaining their own kit – at considerable expense and in all likelihood not having the necessary in-house expertise to manage all but the most basic of problems – so relying on a third party to be available to carry out emergency repairs/problem solving, but a third party that may well not be that familiar with your infrastructure environment?

Ok, so historically, there were some significant technology issues with outsourcing IT. I remember the first wave of Storage Service Providers (SSPs) who, without exception, sunk under the (then) insoluble problem of providing the necessary speed and bandwidth to make the moving around of large quantities of data a financially attractive proposition. And I’ll never forget listening to a Dutch government official explaining how, prior to the deployment of WAN acceleration technology, he used to go into his remote office location in the morning, open up his main work application, go and have breakfast and, approximately half an hour later, on returning to his desk, this centrally-served application would, hopefully, have opened!

And while we’re talking about outsourcing, it would be difficult to ignore the various horror stories that made the news headlines as various (at times it seemed like all of them!) UK government IT projects came in way over budget and way underperforming, or even being scrapped altogether.

However, in 2017 the outsourcing, or managed services and colocation markets are rather more reliable. I’m not sure why any start-up company would go out and purchase any IT/data centre infrastructure when there are a whole host of reputable and reliable infrastructure suppliers able to offer whatever level of service granularity, performance and security are required.

Furthermore, I would go so far as to say that all those companies who are still serving their company’s IT needs from a loosely titled ‘computer room’ (probably a small, dark cupboard/space that has no other use) would be significantly better off trusting their data centre/IT infrastructure with a colocation and/or managed services expert. Even medium sized and large companies who own something more like a traditional data centre need to sit down and do some serious, honest and complicated sums to understand whether or not it makes more sense to outsource IT and facilities infrastructure, to ensure that their somewhat limited IT expertise resources can concentrate on the part of the technology stack that actually makes the money – the application(s).

Logically, the only obstacle stopping the server-huggers let someone take care of their infrastructure is a mind set – how can we trust these guys to look after us? Well, no one seems to bat an eyelid over trusting third parties with the other utility supplies, so why would/should the IT/data centre utility be any different? A power cut tends to disrupt a business, as does a dead phone line, but no one thinks it’s necessary to own and control these aspects of the enterprise.

Furthermore, by using a colo/managed services provider, your organisation will be benefiting from a potentially massive economy of scale – your two servers, networks and direct attached storage has a high capital cost and you’re only using, maybe, one per cent of their respective capacities (and you certainly can’t afford a UPS); a colo/service provider can consolidate multiple customers and IT assets on the same infrastructure, ensuring that it’s a far more cost-effective resource, being used to something like its full potential.

Ah, but many organisations have that one, vital application that runs on the Noah’s Ark server bought in 1975, and that the company’s oldest employee is uniquely qualified to run and maintain. So, how could we possibly move this across to a colo environment? Clearly, you couldn’t, but if you’re relying on nearly 50 year old technology to keep going indefinitely, and for your pensionable employee to live long enough so that cryogenics will allow you to thaw him/her out every time you have a problem, then it just might be that your business continuity/disaster recovery policy needs re-evaluating. Yet, very often, this ‘unique application’ problem is given as a reason not to go down the colo/managed services route. Normally sensible individuals, who back up all their data in multiple locations and have an extra-resilient data centre/IT infrastructure, can’t see the sheer stupidity of refusing to address a significant, possibly catastrophic single point of failure. Okay, so it might be complicated and costly to have the application re-written and to migrate all the historical data across, but I’m guessing the cost to the business when the original application does finally fail, could be somewhat higher?!

So, colocation and managed services might not be for everybody, but there are very few, if any, ‘special’ situations where it makes more sense for an organisation to keep its data centre and IT infrastructure in-house. And, don’t forget, virtually all of the IT giants who are opening up data centres all over the globe as the managed services market takes off, are doing so in virtually all cases courtesy of taking space in colo facilities. If the IT professionals trust the colos, you have to be a fairly special organisation or individual to bet against them.

 

Topics: Volta News

Subscribe here!

Recent posts