In my experience as the Senior Solutions Architect at Blue Chip, there are many different approaches taken by customers when looking for hosting facilities outside their own company premises. A common thread through working on many deals is that the provider is often selected based purely on price and short term requirements rather than considering the longer term. Over the period of a hosting contract this approach can often cost a company unanticipated additional expense, reduced flexibility or both. This blog outlines a few considerations to demonstrate this view.
Firstly let’s look at price. Obviously this is important but across the marketplace there is a huge spectrum in the quality of data centres being offered. These range from Tier 1 premises equivalent to the computer rooms already in use by many companies to Tier 4 where the facility should be truly resilient and have no single point of failure, with most being somewhere in the middle. The table below provides a simple overview of critical parts of the tiering classification. You should expect to pay a premium as you move up the tiering levels, reflecting the data centre provider’s investment.
|Tier 1||Tier 2||Tier 3||Tier 4|
|Number of Delivery Paths||Only 1||Only 1||1 Active1 Passive||2 Active|
|Redundant Components||N||N+1||N+1||2(N+1) or S + S|
|Raised Floor Height||12”||18”||30-36”||30-36”|
|Floor loading pounds/ft2||85||100||150||150+|
|Annual IT downtime due to site||28.8 hours||22.0 hours||1.6 hours||0.4 hours|
|Single Points of Failure||Many + Human Error||Many + Human Error||Some + Human Error||None + Human Error|
|Fault tolerance to Worst Event||None||None||None||Yes|
|Staffing||None||1 Shift||1+ Shifts|
Data Sourced from Uptime Institute, S= System Distribution Path
When selecting a potential data centre you need to ensure that it resolves any weaknesses in the current facility, and will meet or exceed future requirements for service levels such as availability. Just knowing the price per rack is not sufficient. Remember that each time you relocate your company’s infrastructure it introduces risk and outages to services, as well as cost. Additionally, whatever industry sector you operate in, the drive is for increased availability of services, so will the data centre you choose cater for that future business demand even though you don’t need it now?
Communications is also a major consideration. Whilst the cost of dedicated links between locations has reduced in recent years as the demand for ever more bandwidth has increased, the expense related to connectivity can still be significant, not just for initial installation but also whenever a link needs to be moved. The range and flexibility of communication options available from a data centre provider should always be taken into account. Do they offer resilient scalable links from different providers for Internet, can they cost effectively provide a connection into your existing wide network or is this down to you? If the latter, have you considered this cost in the business case?
Whilst reduced communications costs also means that data centres no longer need to be near the companies using their services, this geographic flexibility introduces its own challenges. If you are based in the north of England and find a data centre in the south that fulfils all the other needs we have discussed, how do you deal with the need for checking for loose cables, starting manual reboots, performing firmware and other software upgrades that cannot be done remotely? Travelling from one end of the country is costly both in terms of people’s time and travel expense, let alone the impact on the carbon footprint. Understanding what on-site support is available is essential, not just for basic operations, but also for skilled technical consultancy, and don’t forget there are 24 hours in a day. Is there always someone on-site to help or in the last resort to let one of your own resources on site?
As discussed earlier, it is critical when reviewing data centre outsourcing options and providers to look at the longer term as well as the short term. Where does your company see IT management in the future? Is it a core service within your business? Could your technical resources be better utilised by focussing on the differentiating elements of your IT infrastructure such as the applications and how they are used? If any of these areas are likely to change over the coming years then it is important to ask the question as to whether you just want a data centre provider or a partner that can also provide managed services across the common layers of your IT infrastructure such as the hardware, operating systems, anti-virus, patching, and perhaps database layers.
If you host your infrastructure in a data centre and receive a managed service delivered remotely then the service provider will have some of the same challenges outlined previously around on-site resource provision and activities, all of which can limit the type of service available. This blog has touched on just a few considerations around selecting a hosting provider but there are many more to talk about on another day such as whether a provider’s facility is owned or leased, how is power charged, will price increases be passed on to you, are you considering cloud or Infrastructure as a Service (IaaS), do you need multiple data centres, does the provider meet your environmental policy, what about disaster recovery services, etc?
Blue Chip owns and operates two data centres in the Bedford area, and believes that as a comprehensive service provider it has solutions and offerings to meet the majority of common needs and a few of the exceptions. I can be contacted via our web form or by leaving a comment below. What questions do you need answered?