European DC operators face a twofold threat: The rising wholesale cost of electricity as well as the threat of blackouts. On premise datacenters tend not to have the same level of consumption as larger co-location providers, therefore their negotiating position is weaker. Some DC operators have already negotiated a fixed price for power before the Russian invasion of the Ukraine, therefore they can potentially be sheltered for a lot of the price rises, however most co-location deals over 100Kw consumption offer a pass through price from the utility suppliers, meaning those customers will face the impact. Already I’ve seen price rises between 200 and 300%, with little confidence in the short term that they will drop without some form of government intervention. The UK’s move to cap the energy prices to consumers and business should shield these price rises for 2 years and 6 months respectively. Which removes some of the pain, albeit in the short term. We are watching the EU governments to see if they make similar caps.
The availability of electricity supply from fossil fuels is more of a concern. Some customers I have spoken to are already planning for power blackouts. Most DCs will have some form of redundancy with UPS and generators that will be able to guarantee the continuity of operation, however that depends on the quantity of fuel that is stored on premises. Some diesel supplies may be contaminated by the diesel bug, which, if left untreated, can render some of the fuel non combustible. It is agreed by fuel companies that if left for more than 6 – 12 months, this will represent a very real problem. A further issue is when it comes to re-supply, some of the DCs in the central hub of Paris, Amsterdam and Frankfurt might be unreliable due to the sudden demand affecting all DCs in the area.
Therefore Gartner recommends the following:
1. Test the failover procedures for energy supply from mains power to self generated. Weekly generator tests will not be sufficient to ensure reliable failover. Instead look to a full “black building” test, something that will cause a lot of planning and failover testing.
2. Check the current onsite fuel quality to ensure it has not been contaminated
3. Check the re-supply contracts to ensure there are mechanisms in place for continuity. Also understand the physical location of the fuel depot in relation to your DC, plan for congestion, bottlenecks and failure of delivery – have redundancy in your supplier contract.
4. Examine the utilization of your current server estate. Are there any opportunities to consolidate or decommission equipment?
5. Get visibility of your energy costs. In some cases this isn’t seen as part of the IT budget, falling instead under facilities
6. Audit the failover processes with your co-location provider. Are they doing all of the points above?
7. Currently the cloud providers aren’t passing on any of the price rises. See if there is any opportunity to expedite moving workloads to the public cloud?