How can a data centre’s design contribute to optimum performance?

How can a data centre’s design contribute to optimum performance?

There are a number of elements that contribute to a data centre’s design, including location, efficiency, layout, mechanical design, to name a few. All of these factors contribute to its operability and performance capability.

Mark Fenton, Product Manager at Future Facilities, defines a data centre’s performance as being a mixture of many things and that each business will have different KPIs for their data centre. “Whether the overarching goal is being green and efficient, cramming as much capacity as possible into the space, or to be as resilient to failure as possible; the key to achieving it is in the design and redesign, of the data centre.

“The data centre is a complicated system with many moving parts; power, cooling, racks, IT, cabling and more. Usually, these are independently designed, assembled into one working system. Essentially, it means we flip a switch and hope it will work. However, it often doesn’t.”

Fenton went on to address the independent research commissioned by Future Facilities which found that one third of data centres see their facility’s temperature managed using rule of thumb and that 40% then suffered outages in their data centre because of human error. “Data centre operators must remember that when the parts of a data centre are assembled, they will have a significant impact on one another. This can be no place for guesswork.

“For example, to have a well designed white space, but a poor infrastructure layout on the roof will only result in bad data centre performance. Equally, if you design an efficient rooftop configuration for the cooling infrastructure but don’t consider the white space layout, the end result is also likely to be poor performance. Data centre operators need to consider the complete solution. Luckily, there is a host of technologies available to enable data-driven decisions which will improve the design of the data centre and its operational performance. One such technology is the digital twin. With this physics-based simulation, operators can prototype how those individual elements will operate together and ensure that they get a data centre that is more than the sum of its parts.”

We hear from various experts who continue the conversation in regard to how a data centre’s design contributes to optimum performance.

Wendy Torell, Senior Research Analyst, Science Center, Secure Power Division, Schneider Electric: “First, it’s important to understand what is meant by ‘optimum performance’. For most data centres, availability is a key performance criterion, since the IT loads they support are mission-critical and the cost of downtime is high. However, there are other performance metrics to consider as well – for instance, how energy efficient the data centre is, what its Power Usage Effectiveness (PUE) rating is, how cost effective it is in terms of CapEx, OpEx, or total cost of ownership (TCO) and how sustainable or ‘green’ the design is.

“It’s important to begin the design planning with discussions around all the key criteria for the data centre’s performance to avoid having to go back to the drawing board and get the project implemented on time.

“Key design attributes that can help achieve optimal performance in these various ways include:

  • Efficient design – Operators should select power systems (i.e. UPSs, PDUs) that are designed for high efficiency; and choose cooling systems that allow for maximum economiser cooling hours within their geography. They should also pay close attention to design choices that impact air distribution (containment systems, raised floors, etc.), as those have a big impact on overall energy efficiency of the data centre.
  • Connected systems – Selecting systems that utilise cloud-based, mobile-friendly data centre infrastructure management (DCIM) software, which can be managed remotely and is connected to a data lake, allows the user to leverage information beyond that of their data centre for more predictive maintenance vs. traditional calendar-based maintenance. This helps to maximise the life of components, while ensuring availability since components like batteries can be replaced just before they fail.
  • Redundant systems – If high availability is necessary, operators should design with redundancy in mind; this ensures failures of components or required maintenance activities don’t bring the critical load down.
  • Scalable designs – Future loads are often unknown, so having a design that can easily scale is an important attribute. Many UPSs are modular, so as the load grows, additional power modules can be added. Also consider the current and future density needs. As chip density grows, can the cooling design keep up? There is an increasing trend towards liquid cooling as chip densities climb above the limits of air-cooled data centres.
  • Standardisation – Using standardised designs helps to ensure predictable performance; unique designs can mean unique problems. Pre-integrated, pre-fabricated systems also reduce data centre deployment time, which may be an important performance criterion.

“Today, there are many resources freely available to help decision-makers understand the trade-offs of data centre performance attributes. For instance, the Schneider Electric Capital Cost tool or PUE tool can help operators to see the drivers of cost and efficiency, enabling the right choices to be made given performance priorities for the data centre(s).”

David Friend, Founder and CEO, Wasabi: “Organising your data centre for optimum performance has become an increasingly important consideration as data continues to proliferate at enormous scale, proving integral to business success and local data protection requirements are front of mind. 

“Some data centre operators have started to provide both compute and storage infrastructure as-a-Service. As a result of these design developments, in the near future, data centres might need a specialised ‘storage area’ that is segregated from the rest of the cages for security and cooling reasons. This is because general storage uses less power than compute so has lower cooling requirements. Getting cooling right is paramount as you can easily waste cooling power unnecessarily. 

“There are also ways that data centre cooling can be tailored to large storage areas more efficiently than if storage is spread out among every tenant’s cages – for example, when designing a floor plan of this nature, Compute-as-a-Service and Storage-as-a-Service would likely be walled off.

“Another major consideration when designing a data centre is the size and relative spread of the data centres themselves. Given most data is generated and consumed locally, moving it across long distances between huge facilities isn’t intuitive. 

“At Wasabi, we’re actively considering using a tight network of smaller cloud storage units, rather than a handful of very large storage facilities – the issue is making this work without losing efficiency at scale. While there isn’t ‘intelligence’ on how to make this work in practice, in theory having a tight network of smaller storage facilities that are close to customers, close to the compute, aware of each other and even share metadata, is possible. What this would mean in practice is local data could be cached and accessed locally across the world. The need for this kind of storage architecture will be greatly increased by the rollout of 5G and would have huge benefits for many industries. In healthcare for example, a patient’s records could be accessed instantaneously by doctors in different regions because the metadata is replicated globally, with every storage node able to function as an ‘origin’ server. 

“There are numerous ways you can ensure data centre design contributes to optimum performance, but the cloud industry is evolving rapidly and so must data centre design strategy if it is to maintain and improve upon efficiencies going forward.”

Philippe Ouimette, Director, Strategic Partnerships, Genetec: “My advice would be not to underestimate how your security measures can drive enhanced business performance and customer experience when you are looking at designing a data centre. Done correctly, it can power a frictionless experience for customers (particularly important during COVID-19) as well as making operations much more cost-effective and boosting the employee experience.

“This is best explained using one of our real-world clients as an example. NEXTDC, one of Australia’s most trusted data centre providers, is a great case study in how to use physical security systems to move beyond pure security to improve business operations and enhance customer experience in a highly secure environment.

“NEXTDC had long been using Genetec Security Center to manage access control and video surveillance in its data centres (including video access control intercoms). With its data centre footprint growing rapidly, in 2019 it implemented Security Center Federation to enable centralised monitoring, reporting and alarm management across all nine data centre locations.

“This has been a game-changer as it has unified NEXTDC’s entire national footprint and allows for redundancy and collaboration between all sites.  NEXTDC now has complete transparency and unified control in monitoring facilities effectively right across the country, to protect against and prevent security breaches and ensure a secure environment.

“Genetec developed a highly scalable architecture that synchronises cameras, doors and other managed hardware, access points and solutions. Role-based protocols determined by NEXTDC give customers and personnel appropriate access to video surveillance systems, video analytics and reporting, access control, two factor authentication systems and intercoms.

Jeroen Schlosser, Managing Director, Equinix, MENA: “Digital advances in product design are creating new opportunities to boost both environmental sustainability and profit margins. By making product development more design- and information-intensive, companies are creating a new generation of digitally enhanced offerings that are reducing material use, lowering energy demands and increasing revenues.

“Even as digitisation has accelerated, overall, data centres have flattened the curve on total data centre sector energy consumption, due to a combination of server efficiency improvements, data centre consolidation and individual site design and operational innovations. Data centres are enabling digitisation of the modern world which is driving the consumption of products with much less environmental footprint than their physical or analogue counterparts, as well driving societal value.

“Indeed, data centres are supporting the digitisation of the modern world, which is driving more sustainable outcomes. While it is not measurable as a direct data centre benefit, the carbon footprint of digital products and services is much less than their physical counterparts. For example, an email has a smaller footprint than mailing a letter and downloading an album is more energy efficient than buying a disc.

“Large data centre operators (be they colocation, managed services, networks or cloud) have the economies of scale to apply best practices and drive energy efficiencies across their footprint much faster than small, individual enterprise data centres could achieve on their own. Larger scale also means greater purchasing power when it comes to advancing renewable energy.

“Equinix has been committed to adding value to the business by helping to green its customers and partners’ supply chain, minimising their environmental impact even during these trying times. Equinix global reach, scale and buying power have enabled Equinix to:

  • Invest over US$100 million in energy efficiency upgrades, retrofits and improvements since 2011, which helped avoid 1.1 million megawatts (MWh) and 430,000 metric tons in carbon emissions over the same time period.
  • Purchase renewable energy for 160+ (out of 200+) Equinix International Business Exchange (IBX) data centres.
  • Cover 90% of Equinix worldwide energy consumption with renewable energy. In EMEA, we’ve covered over 95%.
  • Deploy cutting-edge innovations in our data centres that benefit all our customers such as Machine Learning to identify new energy efficiencies. IBX SmartView, our real-time data centre infrastructure monitoring (DCIM) software, helps customers make better decisions to improve their energy efficiency.

“All of these innovations are passed to customers, enabling them to become greener faster and more efficient than they could on their own.”

Martijn Imming, Business Development Manager Data Centres EMEA, Eaton: “Current methods of data centre design and build often involve a ‘mix and match’ approach which assesses components piece-by-piece to complete the power chain. This product approach has been deployed successfully over the last three decades, but cannot deliver the ever growing expectations of the next generation of customers and the data centres they require. Faster construction programmes and leaner builds will require a more integrated systems approach that can reduce cost and improve uptime.

“Hyperscalers have accelerated the evolution of data centres with their fast-paced construction programmes, and where previously, the largest colocation providers could deliver just a few data halls in a year, hyperscalers are now delivering this in a month. By using a single manufacturer, component selection can be finely tuned and more closely matched, resulting in downsizing of the components in the power chain. By adopting a systems approach to the design of the electrical system, many of the problems faced by mix-and-match design can be easily and quickly addressed.

“Such an approach involves considering the design as not just a number of functional blocks, but as a group of intelligent components coordinated with each other and tuned to the act as part of a coherent whole. Modular build data centres maximise the benefits of a systems approach, particularly when constraints exist which hamper historical stick-build construction methods. They are partially or wholly pre-fabricated off-site and provide many benefits as well as optimised performance, including quicker speed of deployment, design and installation cost savings and reduced total cost of ownership. A systems approach to design will define the future of data centres. Manufacturers are ideally positioned to combine complete component understanding with deep-dive design experience to produce a more finely tuned electrical system. This results in not just performance gains for the customer, but also significant cost savings across the complete cycle of the data centre, from the initial build to end of life.

“Smart Uninterruptible Power Supply (UPS) installations can play a key role in a smart systems approach to data centre design. Being integrated with power infrastructure which is more aware of the energy it is using, a smart UPS can more effectively respond to outages, where even a few seconds can be critical, while also providing significant energy savings through efficient power performance. Smart installations combined with a systems approach to design can have a truly transformative effect on the performance and efficacy of new data centre deployments.”

Sarah Doherty, Product Manager at iland: “Currently, businesses thrive on data which is growing at exponential amounts. Therefore, data backup is the initial step needed for protection so that businesses can guarantee smooth operations. It is also used to protect the organisation’s data against damage, cybercrime, theft, software failure and disaster. Data centres are more than just a safe and secure facility with space that is equipped with reliable power and network, they are becoming a valuable addition to many businesses as they prove to be a dependable extension of their IT team. 

“Data centres are limitless entities. The more processing power and bandwith that is made available, the more resources it will demand. Every improvement in a data centre performance is quickly consumed by the latest applications or a new project that requires even more throughput, faster processing speeds and more data storage. Keeping pace is a constant challenge for IT teams. The data centre needs to offer optimal performance, minimise waste and provide for cost-effective operations. Issues like ageing infrastructure, an unsuitable location or poor data centre design can obstruct optimal performance targets.

“Business owners are in constant need of reliability when it comes to maintaining a dependable operation. A data centre is built to withstand a 24/7 service easily, however, the components require a significant amount of infrastructure support in both hardware and software areas. These include power subsystems, stable and uninterruptible power supplies, proper ventilation, high-quality cooling systems, fire suppression, reliable backup generators and connections to external networks. The business world is moving at an incredibly fast pace while matching the overwhelming demand for information with excessive amounts of data creation. With the ever-changing requirements of the modern business model, many companies place their confidence in cloud data centres, as these facilities play a crucial role in reaching their IT requirements. Using the cloud has many benefits over using on-prem data centres. For example, on-prem data centres require a substantial capital investment in hardware. With the cloud, an organisation pays only for the portion of hardware that they use. On-prem data centres also incur added costs for power, environmental controls (HVAC), regular upgrades and tests, security, etc. With a cloud provider, these costs are built into the subscription fee structure.

“Data centre design is a large-scale process. It not only requires a common-sense approach to planning and organisation, but also the help of specialists to guarantee ongoing optimisation within the operation. With expert assistance, a business can create successful strategies, properly manage resources and avoid pitfalls that compromise efficiency and cost money.  With the right cloud data centre design, the business can scale and grow as needed, while still providing the optimum IT performance that is required.”

Gerard Thibault, CTO, Kao Data: “Optimising data centre performance, be that through energy efficiency, security or daily operations, is a continuous process and evaluates operational feedback from improved designs. Key drivers for optimisation are reducing operating expenditure (OpEx) and initial capital cost, which are focal points of discussion for customers’ procurement teams and funders.

“Computing demands are also changing; AI applications are driving up power densities and data transfer, while pre-built, homogeneous rack deployments are driving ROI as hyperscalers expand. As server installations become more power-hungry, ultra-efficient cooling and electrical systems can help minimise customer OpEx, reducing overall energy consumption. With communications latency tolerance becoming more acute and applications under greater pressure to perform, data centres also need to be built in locations that benefit from outstanding fibre access to offer optimum performance.

“With evermore complex RFP processes and greater scrutiny of efficiency and sustainability goals, data centres must operate in a leaner, greener fashion, keeping energy efficiency high, working towards Net Zero and potentially negative carbon emissions. By utilising leading components and learning from large-scale data centres, highly efficient facilities can be built in the UK, offering the benefits of renewable energy and reduced data transport costs to enterprise, cloud and wholesale colocation users.

“PUE has long been considered as the holy grail of performance; however, linking energy to IT performance could be a future metric. Until then, PUE remains a clear indicator of performance or optimisation.

“Other key elements that drive PUE reduction include the data centre architecture; minimising distribution losses, eliminating internal plant heat loads and redundancy that does not increase system complexity. Design must also consider space within the technology suites to ensure it can host increased densities with air cooled servers, but allow flexibility, so that liquid cooling and hybrid environments can be configured. Air containment is essential and the flexibility to have ‘room neutral’ liquid cooling installations in the same environment offers a great opportunity to develop better strategies for heat re-use. As this debate moves further up the sustainability agenda, the data centre will be a fundamental part of a wider ‘community energy strategy’, improving the overall efficiency of the digital ecosystem. Already, we see rack densities at over 35kW and envisage 46kW as the entry point of liquid cooled IT, rack power connection is the next challenge.”

Click below to share this article

Browse our latest issue

Intelligent CIO Middle East

View Magazine Archive