A functional data centre is essential to any kind of cloud strategy, but how can you ensure yours runs at maximum efficiency? IDC research director Donna Taylor shares 10 key issues IT pros should consider when planning their data centre strategies.
Data centre picture from Shutterstock
Taylor was the opening presenter at Data Centre World in London, which I’m attending this week as part of Lifehacker’s ongoing World Of Servers coverage. Here are her top tips and trends for data centre and cloud deployment in 2013.
Take a broad view. If you want your data centre to work at maximum efficiency, you need to consider every aspect of how it intersects with your business. “Top-performing data centre managers do everything very well,” Taylor said. “They don’t cherry pick the low-hanging fruit, do great at that and hope to phase in the other areas. They focus on everything.”
Measurements relevant to data centre performance cover a broad swathe, including cost control, staff numbers and turnover, meeting key business KPIs, the ability to equip new offices and roll out new services, and the cost of IT as a percentage of revenue. There’s plenty of scope to improve in all those areas: by IDC’s calculations, only 10 per cent of companies utilising data centres (whether their own or provided externally) fall into that category.
Measure staff turnover, but understand what it means. While top-performing data centres do well on a range of metrics, the worst performers do have one thing in common with the best: their level of staff turnover is relatively low. “Bad performers often do badly at everything except staff turnover,” Taylor said.
However, the root causes can be very different. In the case of high-performing centres, staff retention is often the result of proactive measures, while in lesser performers, it may reflect nothing more than staff who are scared to try and seek out other work, Taylor suggested. That’s likely to be even more pronounced in sluggish economies, such as Europe where Taylor is based, compared to a relatively prosperous market like Australia.
Automate, automate, automate. “Automation is vital in all data centre activities,” Taylor said. Not only does this improve efficiency, it also contributes to employee retention.
“Advanced automation is key for IT staff. In the data centre they’re often doing very repetitive tasks and not putting their skill set to more revenue generating activities — that’s boring for high-paid staff.”
[related title=”World Of Servers” tag=”wos2″ items=”7″]BYOD is always a pain. The shift towards bring your own device (BYOD) has an impact on data centre strategies. “You’re always creating data, sending data and receiving data any time you touch any of these devices, and this is creating big strain in the data centre,” Taylor said “BYOD creates a lot of additional data and it creates added pressure. Not only is this creating big spikes, it’s also creating security headaches.”
Focus on disaster recovery and backup. The shift towards relying on cloud services, combined with the growth of BYOD, means that ensuring disaster recovery sites are maintained and that backups are automated and rigorous remains critical. “A lot of these trends we’re seeing are really creating DR concerns,” Taylor said. “Everyone is concerned about their backup windows.”
Be bossy about contracts. If you decide to use an external data centre provider, remember that it’s a competitive space and you can be demanding in your requirements. “A lot of the infra that goes into data centres will be dictated by end users,” Taylor said. “The end user has a lot more power to control the T&Cs contained in contracts, and that’s something everyone needs to exercise.”
Include SSD in your plans. The relatively high cost of SSD meant that its initial role in data centres was for caching, but falling costs (and occasional hard drive supply hiccups) have seen its functions extend. “SSD-only arrays are driving disk-to-disk backup and the cost is coming down,” Taylor said. “This dramatically changes data centre performance. Very small amounts of SSD can displace a great number of disks, and that affects power and footprint and cooling issues.”
Recognise that storage spending is still high. While there is a constant pressure to drive down costs (and constant technology improvements helping out), storage expenditure isn’t necessarily falling. “Storage as a total piece of the IT pie has gotten bigger,” Taylor said, noting that trend was true regardless of budget size.
Focus on management and security when virtualising storage. Storage virtualisation is popular, but you need to approach it correctly. “In storage virtualisation, people are looking for ease of management — pooling that data into one area so that the entire system has a good view of it,” Taylor said.
That’s potentially very useful, but needs careful planning. “Security is very much a concern when you’re talking about virtualised data pools. The number of access points may have shrunk, but you still have the potential for someone to gain access not just to siloed storage, but to all of your storage.
Don’t freak out about lock-in. Apart from the oft-cited issues of latency and security, some businesses resist shifting towards centralised data centres and cloud strategies because they worry about having too much technology tied up with a single provider. “Customers want fewer suppliers but fear lock-in,” Taylor said.
It’s sensible to contemplate these issues, but don’t overestimate their importance. “Converged systems offer so much advantage in terms of performance that should be able to overcome any vendor lock-in concerns.”
Lifehacker’s World Of Servers sees me travelling to conferences around Australia and around the globe in search of fresh insights into how server and infrastructure deployment is changing in the cloud era. This week, I’m in London for Data Centre World, paying particular attention to how to maximise efficiency and lower costs in the data centre.