Jonathan Healey, director of Systems Engineering & Technology at Schneider Electric, explores the complexities of distributed data centres at the edge and the role cloud-based software has to play in ensuring availability and reliability.
The rapid growth of data shows no signs of slowing. Smart-devices and emerging technology trends enabled by the Internet of Things (IoT), in addition to high-quality bandwith-intensive content and intricate new concepts such as driverless cars are driving the volume of network traffic to new heights.
The number of network-accessible devices including smart phones, tablets, Point-of-Sale (POS) systems and other forms of wearable-tech are also continuing to increase. Cisco expects that by 2020 there will be as many as 50 billion network attached devices globally, and within the same time-frame, a further 20.8 billion IoT-enabled devices will also be connected, which will help drive a threefold increase in global data centre traffic during the next five years.
It is expected that annual global data centre traffic will reach 15.3zettabytes (Zbyte) by the end of 2020, increasing from 4.7Zbyte just three years ago in 2015. The surmounting challenge of Big Data is driving change within the data centre industry, in many cases via the proliferation of smaller, micro data centres at the edge of the network.
The edge; by definition
According to the Infrastructure Masons: “An Edge location is a computing enclosure/space/facility geographically dispersed to be physically closer to the point of origin of data or a user base.” For many of today’s businesses, moving data centres to the edge has become a matter of necessity: as network-enabled applications become more critical, the latency, or speed of reaction of IT infrastructure to data traffic, requires that local compute must be placed in closer proximity to where the data is needed or consumed.
Larger, centralised facilities remain an important factor in the data-centre ecosystem. However, they are more commonly used for hosting slower applications or housing archived information, providing services for e-mail and other such functions, for which speed of response is not so mission-critical. But the data-centre pendulum is swinging away from the centre and towards the edge of the network.
The evolution of the edge
There have always been small data centres distributed at the edge of networks, running applications for small businesses (SMB’s), Small-to-Medium Enterprises (SME’s) or remote office branches of larger organisations.
Over the last 10 years, many of those data centres and their applications have been consolidated or replaced by the cloud. Now, given renewed focus on the edge, enterprises are experiencing the difficulties of managing a hybrid cloud/edge architecture, whilst ensuring performance, security and availability throughout
Fortunately, data centre technology is evolving to meet these challenges. An increasing trend has seen standardisation of the computing, networking and storage infrastructure found in these facilities, which allows customers to choose the best equipment from different vendors to meet their specific needs. Standardisation also helps drive down costs so the capital expenditure required to build and deploy new edge data centres is thereby reduced.
Technological improvements in infrastructure equipment such as power and cooling systems help to reduce operating costs while providing ever greater reliability and uptime. Schneider Electric maintains that its products, which are designed for ultra-efficient operation, have achieved an 80% improvement in efficiency due to developments in data centre power and cooling technology.
Inevitably however, the monitoring and management of a larger number of distributed data centres presents a unique challenge.
The complexities of a distributed edge environment
By their very nature, many edge data centres are found in facilities that lack the security, support and specialist expertise they need to be maintained. A great advantage of larger data centres is that technical experts will often be located on site. For smaller, distributed data centres however, this is rarely an option.
This in turn may raise the threat of potential vulnerabilities within the networks used to manage power and cooling equipment being exploited to gain access to data. A recently documented story tells of a US retail company where hackers gained access to corporate information by exploiting weaknesses in the cooling management system.
It is therefore necessary that data centres at the edge not only follow appropriate security practices, which ensure that they are physically secure, but that they incorporate safeguards against virtual intrusion into their systems. This would include physical separation between corporate and infrastructural networks so that critical financial or customer data is never compromised.
Fortunately, for many of today’s businesses embracing the age of digital transformation, software management tools are also evolving to meet this need. Some of the most innovative and market-leading software makes use of the very cloud infrastructure data centres provide to deliver unique data insights, monitoring and management capabilities to ensure uptime and resilience.
Software holds the key
Software applications such as Schneider Electric’s EcoStruxure IT, Data Centre Management as-a-Service (DmaaS), part of the EcoStruxure for Data Centres architecture, offer a vendor-neutral, efficient and reliable method to managing, maintaining and identifying any potential troublesome requirements of data centres at the edge.
Through DmaaS, network-enabled embedded sensors gather data on the status of individual hardware products within the data centre and transmit the information to a centralised network operations centre (NOC), which may be located at the premises but more likely at another remote site.
There the data is collected and analysed for service and maintenance purposes, or to respond to any emergencies identified. Furthermore, the information can be pooled and analysed to provide a machine-learning function with the ability to predict when a product is likely to experience an unplanned issue or potential failure.
Detailed data, including maintenance schedules or emergency alerts will then be transmitted remotely to local experts, direct employees or independent service providers operating to predetermined service level agreements (SLA’s) to ensure that maintenance, if required, is dealt with faster and more efficiently than previously possible. Using such software tools in conjunction with mobile applications and specialist partners, will enable the similar levels of security and reliability normally associated with larger data centres, to be delivered to smaller, distributed edge facilities.
The right software can also increase efficiency by improving staff productivity, providing greater maintenance coverage with fewer personnel and deliver more widespread and more accurate visibility of all IT and infrastructure equipment connected to the network; resulting in faster response times and greater uptime for the owner or operator.
Conclusion
By harnessing the power of cloud-based management software systems and innovative, connected infrastructure architectures, distributed edge data centres can wield the power of the IoT to deliver exceptional cost-savings via resilient digital services. They can also provide a platform for growth with faster more-reliable, lower latency internet speeds that ensure businesses can truly adapt to an ever-more digitised, consumer environment.
This comparison of small edge data-centres vs. centralised cloud data centres truly is a story of David and Goliath. Both types of data-centres are necessary to sustain the application performance, security, and availability of the enterprise applications companies of all sizes need. Schneider Electric’s advancements of power and cooling technology, as well as DMaaS, ensures both owners and operators can build and run their overall IT architecture cost-effectively and efficiently from the cloud to the edge.
