In their attempts to achieve higher levels of operational efficiencies, enterprises across all verticals are rapidly building their data centre capabilities for efficient storage, management and dissemination of data. However, with the growing adoption of next-generation technologies such as the internet of things (IoT) and artificial intelligence, there is a perceptible shift away from centralised data management to Edge computing, which enables organisations to process data as close as possible to end users. With the help of Edge computing, organisations are moving part of their data processing to the periphery to improve network performance, decrease data traffic and reduce latency.
Key factors driving adoption
Over the past few years, the demand for Edge networks has been on the rise, driven by the need for real-time data processing. According to industry estimates, by 2025, almost 20 per cent of the world’s data would have to be analysed in real time, closer to end users, rather than being sent to core networks for processing. Edge computing can help businesses to reduce the need for transmitting data back and forth across networks, thereby reducing bandwidth costs. From customers’ perspective, edge computing can help simplify and localise the delivery of applications, data sets and services.
Edge computing is particularly useful in industries where high latency is an issue such as retail stores where unstable network connections significantly impact the customer experience and autonomous vehicles that require low latency and close proximity to users. In the healthcare industry, edge computing can help medical professionals gain real-time access to patient data, thereby making it possible to improve diagnosis and patient health monitoring over long periods of time.
In industrial manufacturing, edge computing allows manufacturers to gather data and gain insights into predictive maintenance and energy efficiency.
Much of the shift to edge computing can be attributed to the growing adoption of IoT. In an IoT architecture, there are millions of end-point devices that produce large amounts of data, and this data must be processed and acted upon to draw meaningful insights. This, therefore, requires sending lots of data through narrow pipes to data centres. Since the cloud providers charge by the amount of data that is being processed, it is in the enterprises’ financial interest to reduce the amount of data that they send up for processing. In this context, edge computing obviates the need to send huge volumes of data to central data centres by doing much of the processing work at the edge of the network. Edge computing, therefore, increases the performance of applications and relieves increased bandwidth requirements from core networks. Some estimates suggest that edge computing can potentially provide improved latency and data transfer reduction to the cloud of up to 95 per cent. The adoption of edge networks will continue to grow with the increasing proliferation of IoT-enabled devices. In fact, according to the International Data Corporation, in the next three years, around 45 per cent of IoT-generated data will be stored, processed, analysed and acted upon close to, or at the edge of, the network, and over 6 billion devices will be connected to the edge computing solution.
Moreover, enterprises are increasingly adopting edge computing to solve the challenges associated with managing a cloud infrastructure. By transforming cloud computing into a more distributed computing cloud architecture, edge computing ensures that disruptions are limited to only one point in the network. For instance, in case of a natural disaster or a cyberattack, the impact would be limited to the edge computing device and the local applications on that device rather than the entire network.
Impact on data centre architecture
As enterprises increasingly add edge elements to their core networks, data centre architecture is likely to witness a major overhaul, particularly in terms of size and location. With the growing uptake of edge computing, enterprises are expected to gravitate towards a more distributed data centre infrastructure, setting up a large number of smaller data centres near highly populated areas such as cities as well as deploying micro data centres at the base of telecom towers and other important nodes in the existing wireless networks. Therefore, there is likely to be a far higher number of data centres in the future, but these would be much smaller than the warehouse-sized data centres that exist today.
Micro data centres, deployed at telecom tower sites, have emerged as a technology of choice for several enterprises as they help them meet the growing computing needs across edge environments. The key advantage of a micro data centre is its flexibility that can be used in a wide variety of applications. Moreover, the deployment of micro data centre solutions is quick and easily repeatable, and does not need sophisticated IT support. Most micro data centre solutions that are being offered today are preconfigured, pretested, and pre-integrated, providing hyper-converged computing, environmental monitoring and power distribution, all in a secure rack. Micro data centres also require very little physical space and the time to configure is minimal. As many remote locations do not have on-site IT staff, several vendors are offering simpler plug-and-play micro data centre solutions that require local sites to only provide power and communication connection to turn the system on. These tools help remote administrators to monitor critical micro data centre performance details such as available backup battery runtime, and ensure the availability of round-the-clock data centre systems, regardless of the remote location. The tools also enable predictive maintenance, which helps in avoiding unplanned downtime.
Implications for security and storage
The addition of edge computing to core networks poses a number of cybersecurity challenges, which the regular data centre operators might not be equipped to deal with. The likely increase in the number of smaller data centres closer to end users would result in enterprises demanding a replication of the performance and security which they get in centralised data centres, in distributed data centres as well. Industry experts contend that going forward, enterprises will be faced with a tremendous data centre infrastructure management challenge to make sure that the distributed IT facilities are secure. To this end, it would become imperative for enterprises to choose security solutions at the start of a planning process and not at the end once new models are ready to be rolled out.
Meanwhile, given the fact that edge computing by design is autonomous and not an extension of an application running at the core or subservient to the public cloud, the data storage requirements of enterprises are also expected to grow considerably, both in physical terms as well as in terms of virtual storage. The data will also need to be made more local and readily accessible.
Recent technological breakthroughs have given enterprises the flexibility to decide whether their applications are more efficient residing in the cloud, within a traditional data centre, or on the network edge. For the past several years, large traditional data centres have been the mainstay of computing and connectivity networks. Essentially, processing of all transactions has been carried out in a centralised core, but going forward, the burgeoning demand for real-time processing means that businesses will increasingly add edge elements to this essential core. Edge computing brings the network closer to users and eliminates latency issues. It also helps streamline the flow of traffic from smart devices within IoT systems.
Industry estimates suggest that while 91 per cent of data today is created and processed inside centralised data centres, by 2022, about 75 per cent of data will have to be processed at the edge. The introduction of 5G networks will accelerate the trend of data centres adding edge networks, extending their reach closer to end users than ever before. This, however, doesn’t mean that the huge data centre model that exists currently will become obsolete in the future. The edge-driven systems will work alongside the current data centre model and the industry is poised to witness a larger number of smaller data centres built closer to population centres.