Energy Costs Of Data Processing

Energy Costs Of Data Processing

The rapid increase in energy costs will impact on those relying on data storage and processing

Data Processing is moving from local servers to remote data centres offering cloud services. The ease of expandability and what have been relatively cheap running costs has led to increasing volumes of data being processed remotely. Unfortunately the whole infrastructure is still highly dependent on a reliable power supply.  In the UK consumers tended to rely on longer term contracts that guaranteed tariffs.  These are now expiring but suppliers are not offering alternatives rather changing rates as global costs increase.  At best we will see these costs passed on as increased processing and storage charges at worst suppliers could go out of business taking the data they store with them.

Power costs are an important consideration for data centre operators.  There are costs to run the server units together with those required to keep those units cool.  Data processing requirements include tasks received from customers as well as the algorithms and operating systems needed to keep the whole show on the road.  The heat generated by the servers is usually wasted and requires cooling systems to keep them working.  The UK and most of Europe is usually cool enough to use ambient air for cooling rather than requiring additional refrigeration.  Unfortunately the heat wave of July 2022 caused some UK centres including those used by Google and Oracle to fail.  Increasing Summer temperatures are forcing data centres to invest more on cooling.  Some are being built underwater in an attempt to match proximity to users with ease of temperature control.

Energy savings can be made by siting data centres close to sources such as Hydro Electric plants or in countries where energy costs are lower.   Sometimes the data hub needs to be close to the centre of activity such stock trading where data latency needs to be very short.  Connection speeds may not be fit for purpose in remote, poorly developed areas.  There is also the issue that governance and legislation in some parts of the world may not offer sufficient protection to data and its lawful use.

This all means that many data centres will be unable to avoid increasing energy costs and that it will lead to higher processing costs for the consumer.  There may be initiatives from management to keep costs down and these need to be weighed against the security needs of the network.

With many tasks run out of house as ‘software as a service’ cost savings might be made from moving supplier but this will probably involve new training and work routines as another service provides a similar model but in an entirely different way.  A service provider might host their own data or outsource it again to a major supplier such as Amazon.  It is the ultimate data holder that will be paying to host and run the servers.  As energy costs are now such an important factor the actual physical location of data stores should be investigated when planning to adopt or change a ‘software as a service’ provider.

Power outages might become more frequent as the supply of fuel to generate power is constrained.  Outages can be overcome by local generators and batteries if a temporary issue. If a data centre is in financial difficulties the outage could become permanent.  Dealing with such issues should already be part of an organisation’s disaster recovery plan. It is not enough to trust that any outsourced provider will sort it out.  All possible data should be in off-line backups.  It is a long time since that was a case of storing a few tapes or disks.  A recovery plan needs to consider how much data is in ‘cold storage’ off-line, the time taken to restore it and what the business will do until a system is restored.

Kindus undertake cost benefit analysis (including factors relating to energy costs) on SaaS (Software as a Service) providers for our clients.   We are also experts in data loss prevention and disaster recovery.  Kindus is here to help shield against the consequences of increasing energy costs.