Peter Westwood, data centre director, SPIE UK, takes a look at energy efficiency within the data centre, and how addressing capacity planning challenges could be the key to a more environmentally sustainable, cost-effective facility.
The growth of data generation on a global scale is proving to be a driver of increased data centre energy consumption. In fact, there are predictions that show a trebling of data centre energy consumption in the next few years. With this in mind, providers and owner occupier organisations must constantly manage and rationalise their power strategies to meet the increased data volumes.
Traditional data centres didn’t used to experience network spikes because physical servers had excess capacity, often only supporting single applications due to them being regularly sized to manage peak data needs. As a result, normal utilisation was as low as 20%. With virtualisation and cloud computing, servers are highly utilised, being covered by fewer locations, which results in network traffic continually changing.
The Spring 2007 Data Centre Users’ Group survey found over 70% of participants used hot-aisle/cold-aisle configuration to increase cooling capability to respond to higher data demands. The survey findings also revealed that over 60% incorporated air management technology to reduce hot air recirculation.
Providers and organisations are currently working to develop ways to respond to this huge expansion in data traffic. Part of their strategy is to create flexible capacity usage plans. However, these also need to meet service level agreements and avoid risk of downtime or breaching capacity ceilings – both of which are a significant challenge. Put simply, building in spare capacity can increase operating costs.
So how can data centre companies mitigate this issue and create a more efficient facility?
Firstly, proactive planning data centre operators can move towards optimising the capacity of their data centres to meet the needs of IT and business services with just-in-time deployment. Smarter data centre capacity management can help them determine what they need to buy and when.
In the past, operators of data centre operations relied primarily on manual data management to understand how the facility’s requirements were changing over time and growing in capacity and performance. Scaling up would usually mean to overprovision, compared to normal operations which, in turn, increases operational expenditures including power consumption.
Virtualisation can also help with some of the physical issues of managing data centres, using less hardware to handle more workloads. Flash storage is an alternative solution, due to its ability to handle the random input (IO) patterns more efficiently than legacy disk-based storage.
Inefficient data management still contributes to storage waste, even on flash-based systems. But flash-based storage is a way to deal with over provisioning, to reduce energy consumption, creating the storage performance needed but in a smaller footprint. This space utilisation efficiency is also important to operators for improving resilience of existing facilities. It can also assist in reducing power in the data centre, contributing to performance improvements.
Today’s world is rapidly changing, and data transition is at the heart of it. So, what can operators of data centres look to do when seeking demand management improvements and greater operating characteristics?
1. Data Centre Management systems for capacity management can help determine equipment and infrastructure provisions by reviewing the utilisation and existing capacity of rack space, rack power, UPS power, upstream breaker or panel power, cooling, fibre or data port connectivity at the rack, patch panels, and switches. This can determine usage levels of the required components and whether each one of them is fully utilised or has additional capacity. By maximising capacity of space, power, and networking, data centre organisations can make significant resilience and performance improvements.
2. Studying the IT equipment with a view of moving to the new ASHRAE operating standards could greatly assist in reducing energy demands by elevating environmental conditions.
3. Promoting the corporate responsibility and social need for green solutions will contribute towards avoiding the huge predicted increase in data centre energy consumption.
4. Considering different architectures such as Open Compute projects or high-density racks instead of standard cabinets can lead to greater improvements in space utilisation.
5. Applying the latest innovative cooling technology coupled with better energy management arrangements will provide an increase in return on investments.
6. Taking into account the location, annual weather conditions, site constraints, and user requirements will help develop a best fit data centre cooling strategy for the options available.
7. Users can consider the utility service options with associated charges for possible minimisation of peak demands, resulting in cost benefits.
By properly planning ahead, data centre capacity management companies will achieve better efficiencies with their physical infrastructure and data centre configurations. Additionally, it will identify potential risks, anticipate sources of failure, increase operational efficiency, and achieve dynamic intelligent operational management. All of which will significantly contribute to a more environmentally sustainable and cost-effective data centre industry.