FACILITIES capacity
per cabinet, for example, would leave all the power utilised but 80% of the ports unused when deploying a blade chassis, whereas rack mount servers could fill all the ports but leave 50% of the power capacity wasted.
Cooling
Cooling is one of the most important – but least understood – elements of a data centre, not least because air’s invisibility makes its movement around a data centre extremely difficult to track and manage.
To add to the complexity, every data centre technology requires its own method of cooling for optimal performance. Even different generations of the same model can sometimes require different cooling.
The problem is compounded by the failure of either IT or facilities departments to take responsibility for managing airflow in a data centre, with IT controlling equipment deployment but not cooling systems, while facilities strives to remove the total heat build-up but doesn’t influence the deployment of equipment that will change the all-important air flows.
Dr Warner notes: “We can see that it’s impossible to ignore the symbiotic relationships between space, power, network and cooling. How each is managed has a huge effect on the others, making this a highly intricate and sensitive balancing act. However, the different parts of the data centre eco-system must be balanced in this way to extract the full benefit from the investment made in the facility.”
Calculating the cost of lost capacity The Uptime Institute’s 2008 paper A Simple Model for Determining True Total Cost of Ownership for Data Centers provided the industry with a spreadsheet tool for modelling the true total cost of ownership (True TCO) of a data centre project.
Future Facilities’ paper uses this tool to make an estimate of the True TCO for a 1.3MW data centre over a 15-year lifespan, in order to highlight the cost of inadequate capacity planning.
Modelling True TCO takes account of three distinct elements: £ The capital expenditure required for construction and fit-out of the facility
£ Fixed operational expenditure (such as staff costs and taxes) £ Load dependent operational expenditure, predominantly the cost of power and cooling
There is, of course, also the cost of buying the IT equipment installed
in the data centre, but this is not relevant to calculating True TCO since it is dependent on the needs of the business rather than data centre capacity – i.e. the servers must still be bought even if they have to be placed within a different facility.
In the white paper’s example, capital expenditure amounts to $50m, which the True TCO calculator annualises to $6.3m a year for the 15-year lifespan, based on the static annuity method. Fixed costs, which generally fall outside the influence of the data centre owner, amount to $3.1m a year. The major variable is load dependent operational expenditure, because it is determined by the amount of load installed in the facility. However, this is also the smallest cost factor by far.
This graph shows the financial effect of a data centre running under capacity compared with one running at 100%. Even though 50% of the required processing power is not being delivered, the total cost figures are almost identical, due to the minimal impact of savings in load dependent operational expenditure.
To exacerbate the problem, that ‘missing’ 0.65MW of processing must still be delivered, forcing the construction of a second data centre, thus doubling the cost per kW of processing that was originally budgeted for.
Not only that, but any capacity crunch while the new facility is being built could result in delays to the deployment of critical new systems that might potentially cost the company a commercial advantage.
How to avoid lost capacity
As the white paper acknowledges, the pace of technological change means avoiding lost capacity completely is an impossible task. However, it can be minimised by looking more closely at the longer term impacts of each management decision.
To achieve this, data centre owners must improve the flow of information between all the stakeholders in the data centre space, in particular IT and facilities. With more data to hand, modern simulation techniques can be employed to show the effects of future plans within the timescales required for operations.
Future Facilities’ King concludes: “Simulation techniques are the only way for data centre operators to gain a clear picture of the future state of their facility. With the growth of DCIM tools we now have the raw data to work with, while state-of-the-art CFD applications such as our own have the power to generate the modelled evidence to help prevent decisions being taken today that could have major capacity implications tomorrow.”
50
www.dcseurope.info I June 2012
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56