Michael O’Keeffe Issue: Europe II 2016
Article no.: 5
Topic: Physical networks: Innovation and optimisation for the future
Author: Michael O’Keeffe
Title: VP Service, EMEA
Organisation: Emerson Network Power
PDF size: 190KB

About author

Michael O’Keeffe started his career with Chloride in 2003 as UK service director and was promoted to service general manager in 2005. In 2011, following the Emerson acquisition, he was appointed managing director for the market unit based in the UK & Ireland. Since 2014, Michael serves as Vice President Service for Europe, Middle East and Africa (EMEA).
Michael holds a bachelor’s degree in electrical and electronics engineering from North Staffordshire University. He is also a graduate of the Emerson Leadership Program.

Article abstract

puts a strain on the network and our data centres and has led to the emergence of edge computing technology as a form of alternative data centre technology. Edge computing moves key data processing and network services away from the central data centre out to the ‘edge’ of the network. As a physical system, this often consists of multiple small networks linking in to a large central network and data centre. This can result in an enhanced cloud computing experience for the consumer, creating better connectivity between the server and the devices at the edge of the network. However, this is just the latest trend in data centre networking and who’s to say that in a few years the network won’t be built differently? 

Full Article

Looking back at the evolution of infrastructure in the physical network over the last 20-30 years, we have almost come full circle. Before “i-devices” and talk of the Internet of Things (IoT) were even on the horizon, we had the era of the mainframe. Data was managed, processed and stored on one central system. Paper cards, tape, batch jobs and dumb terminals were all day to day necessities and in order to support this equipment, rigid guidelines about power, temperature and humidity had to be adhered to.
Over a period of time, technology advanced and mini computers, what we now call servers, were introduced and we started to move towards a more distributed computing model. As technology progressed, dedicated equipment for different applications, company departments, storage and networking needs, meant more hardware in more places. Power and cooling requirements were still fairly rigid but rather than large central data centres with a central plant, we saw the advent of the “computer room”.
Now, as cloud computing and virtualised systems become increasingly common, we are moving to a more hybrid environment with large centralised data centres, edge computing (to improve speed) as well as significant requirements for data centres in highly regulated industries, although nowadays the terminology and technology is very different.
The rise of edge computing
Our increasing reliance on cloud computing and online services, both as consumers and in our professional lives, puts a strain on the network and our data centres and has led to the emergence of edge computing technology as a form of alternative data centre technology. Edge computing moves key data processing and network services away from the central data centre out to the ‘edge’ of the network. As a physical system, this often consists of multiple small networks linking in to a large central network and data centre. This can result in an enhanced cloud computing experience for the consumer, creating better connectivity between the server and the devices at the edge of the network. However, this is just the latest trend in data centre networking and who’s to say that in a few years the network won’t be built differently?
Data centre construction
Equally, data centres themselves are starting to be constructed differently. Traditionally, data centre building has involved a “bricks & mortar” approach, where data centres evolved organically. This often resulted in a mix of hardware from different providers, meaning a heterogeneous data centre ecosystem. It was also a lengthy process to build a data centre, requiring much planning and project management. Now, prefabricated, modular data centres are coming to the fore, enabling a more flexible and convenient planning and building process. The modules are designed and manufactured off-site and then subsequently delivered and rapidly assembled on-site. This approach means that overall deployment time can be reduced by up to 40%. Furthermore, prefabricated data centres are now far beyond the old, simple industrial designs and have a much wider choice of features and appearance than the standard shipping containers. Modular units can now blend in with their surroundings just like bricks and mortar – a welcomed bonus for many.
The introduction of modular critical infrastructure, power and cooling, also allows for a significantly increased level of flexibility to the data centre industry.
Physical network essentials
Despite these various models, the physical needs of the data centre have remained the same in order to work at the optimum level. For example, the need to ensure critical power continuity has been consistent throughout the different eras. Similarly, keeping equipment within a set temperature range, albeit increasingly elevated, is fundamental to ensuring the high levels of availability demanded.
All of these changes mean that the future is just as unpredictable and poses a challenge to IT and data centre providers. We have had to design and build products, services and solutions for customers looking to invest in a physical infrastructure, which will last a minimum of 15 years, often much longer. This has to be accomplished, with the full knowledge that the technology inside the data centre will often become obsolete in only three years or less. In fact, in some cases, technology refreshes are happening in just 18 months, e.g. social media applications, but for most infrastructure technology, such as power and cooling, it’s fortunately somewhat longer. These rapid advances are taking place within the context of a physical infrastructure that is generally not designed for such accelerated changes.
Indeed, the challenges of the future are only getting more diverse. We are now in an age where the IoT and demand for technology in almost all contexts will widen the gap between technology refreshes and what an aging physical network infrastructure will support even more. IT decision makers therefore need to future proof their physical networks to ensure that their data centres are able to cope with these refreshes.

So what is the solution?
We are already addressing this challenge on a daily basis. A data centre built just five years ago could already be out of date and unsuited to the needs of today’s IT manager. Upgrading and optimising existing infrastructure has never been more important and is a key part of our service business.
Emerson Network Power has a dedicated data centre optimisation team supported by over 3,100 service field engineers. Our team is able to assess and audit every aspect of a network’s critical infrastructure to highlight areas of improvement and propose methodologies for expansion and/or reductions in IT space whilst ensuring optimised operation and often significant energy savings.
Not only can mature infrastructure be easily upgraded but in many cases it must be to meet emerging technologies, as well as to improve on existing efficiency and availability levels. These improvements range from the addition of power and cooling capabilities to simple solutions such as replacing AC fans with more efficient EC fan technology or replacing traditional pumps with permanent magnet motors. These improvements however, are often just one part of the broader optimisation program.
Every infrastructure is unique and therefore its optimisation needs to be thoroughly tailored to the specific site and needs of the IT solution it will support. A telecom data centre will have vastly different future requirements to that of a bank or a hospital. This is why Emerson’s optimisation experts work closely with customers to focus on their specific needs.
Trying to predict the future is always a thankless task and as recent technology advancements have shown, new technology can come on in leaps and bounds and in unexpected ways. We cannot guess what the data centre of 2030 might look like but we have two certainties to reassure us: data centres have a few key requirements, such as constant power and acceptable temperatures, and most importantly, existing equipment can be pushed further than the manufacturers may have envisioned originally. IT managers therefore need to expect the unexpected, but also know that with the right optimisation program they can continue to rely on their current infrastructure for years to come.