Some advice before jumping over the edge of the network cliff
by Blogs & Opinions
2019 was the banner year for 5G hype. Operators across the globe turned on their 5G networks for the first time, and chatter about the possibilities of these networks reached a fever pitch. 5G handsets began hitting the market, to the bemusement of some consumers who questioned the need for the new, heavily marketed standard.
But 5G isn't a consumer proposition – the promise of additional revenue opportunities and optimized operations has been most attractive to operators and enterprises looking to derive value and circumvent long-standing legacy limitations. KPMG estimates that $4.3 trillion will be unlocked across the major industry verticals through the fruition of 5G use cases over the next seven years. New revenue opportunities will be provided via Internet of things (IoT), location tracking, smart cities and even video conferencing (which, although ubiquitous, is workable at best with current 4G networks).
Delivering on unfulfilled potential
The trouble is, 5G-enabled devices – such as connected cars, 5G-enabled systems, smart building interfaces and other low-latency applications – are chatty. They generate high transaction volumes and can produce a huge volume of data. Generally, this volume is too much to be processed at a traditional centralized data center. This means the fate of these use cases, and our ability to deliver them, is linked to how successfully we adopt and implement edge computing.
Enter edge computing – infrastructure facilitating data collection and processing at the network edge. Bringing data to the edge means lower latency and improved user experience while reducing precious WAN bandwidth. It's an attractive proposition for enterprises looking to deliver on 5G's true potential.
Edge computing has been widely discussed for a few years now. It might sound easy, but it isn't. Serious considerations need to be factored in. The progress of the telecom industry is a major dependency, and edge computing represents an unprecedented degree of network outsourcing most enterprises won't be used to.
The drive toward edge computing usually boils down to a need for lower latency. Some industries, such as financial trading environments, require the ability to rapidly store, access and process data. Small e-commerce delays could trigger thousands of dollars in lost revenue and possibly billions for fintech traders. In other cases, specific applications require lower latency. Take medical wearables as an example – latency issues could affect performance and result in a serious emergency.
Enterprises looking to leverage 5G (and therefore transition to edge computing) must clearly define and understand the opportunities it could unlock. This will not only help IT pros make the case internally by educating stakeholders on the need to invest but also provide a useful milestone for determining all the factors necessary to consider.
Like any major networking change, there are costs organizations must account for. Re-engineering applications to ensure they work at the network edge is expensive – as is training staff and making the necessary infrastructure upgrades. Enterprises serious about the edge should be prepared to set sufficient budget aside.
Once enterprises have defined how and why they want to shift applications to the edge, the next step is evaluating what this means for their IT pros' ability to manage key facets of performance and security.
Historically, enterprises have been able to manage every user by monitoring a centralized or small number of data centers facilitating all connections. Running applications at the network edge removes this point of control. Users of your applications will be connected to systems directly rather than to your data center firewall. Without location-based insight into the status of services and applications, understanding user experience at the network edge can be challenging.
Enterprises have already made some difficult choices in entering relationships with public cloud providers. Similar choices apply to enterprises embracing edge capabilities. BT isn't going to share the performance details of their snazzy new 5G network any more than Microsoft Azure or Amazon Web Services are going to give you access to the physical interfaces backing their multitenant platforms.
Adopting network edge services means outsourcing control, which can lead to less visibility into application performance for IT professionals. This is significant, given users' high expectations for performance and a potential skills gap with new technology. If an application or service isn't performing well, users will quickly find an alternative, which presents a major problem for enterprises looking to secure the perimeter.
Maintaining monitoring standards
To counter these performance issues, enterprises will need to make changes to their existing network monitoring practices. Edge computing no doubt makes networks opaquer, but the right solutions can restore much-needed management and visibility.
Application performance management (APM) solutions are the ideal tools for making sense of decentralized networks. APM can reveal performance information specific to users' locations and grant visibility of the individual components applications or services depend on. If this seems familiar, it's because it is. APM techniques are applied cyclically to bridge years-long gaps between new – but unobservable – infrastructures and mature platforms with robust management and monitoring interfaces.
Technological advancements like edge computing require new considerations, but the need to effectively monitor networks remains as important as ever. APM is vital to understanding and actioning user experiences taking place away from the network core. Any application or service moved to the network edge should, therefore, be accompanied by a strong APM solution.
Capitalizing with caution
5G's more fanciful use cases have been hyped for a long time. Emerging edge computing capabilities make these use cases possible, but many remain unfulfilled with tantalizing potential.
5G networks are giving enterprises an opportunity to offer a range of modern, low-bandwidth applications from the network edge. When looking to capitalize from the edge, however, enterprises must monitor performance with the same clarity they always have – albeit with some tweaks. IT pros need to ensure they can maintain a comprehensive view of network performance, despite workload migration to the edge's less observable infrastructure.
With edge computing, network and application performance is everything – it's the differentiator setting competitors apart. APM monitoring tools will help establish more visibility. With its expanded combination of end-user metrics and service delivery component performance, APM is a logical approach to ensuring the performance of modern networks, as it should be. It's always the go-to tool when taming the next big industry trend.