Edge computing dominated MWC 2019 along with 5G and all the robots at the show. In fact, according to some analysts, edge computing could be worth almost $7 billion within the next three years. Much of the new architecture’s advantages stem from the capacity offered by 5G to deploy scalable, typically cloud-based, compute platforms at the edge of the network. However, a growing number of operators are coming across a challenge when they look to scale services to the edge – portability is a headache.
A bad to worse headache
Applying cloud-native principles such as DevOps, Agile or continuous delivery to shift services between devices, local clusters, or the cloud requires a degree of flexibility. Many operators have failed to align their IT needs with appropriately scalable cloud platforms. While some have adopted virtualization, NFV, and Infrastructure as a Service (IaaS) platforms to capitalize on the latest advances in cloud computing technology – business and delivery model for is considerably more complex.
When considering edge platforms, operators and enterprises alike must take into account the technologies commonly used by developers. And it is no secret that developers use cloud compute for applications and deliver them over app stores. The platform must therefore be cloud-native to make the transition for these developers as friction-free as possible. Some operators and enterprises have failed to take this into consideration.
Next, consider the large amount of data that must be stored on edge compute resources. Indeed, the platforms running at the edge must be capable of handling streaming analytics for thousands of devices – with new use cases arising on an almost daily regular basis. Needless to say, interoperability is key and such deployments require a continually evolving ecosystem. This ensures that developers can build, manage and monetize the platforms – now and into the future.
Open source and containerization
A way of meeting both developers’ needs and addressing the challenges faced by operators is Open Source. Why? It has matured over the years and is now seeing greater adoption. For example, the StarlingX project has been designed to bring the developer community closer to the edge and widely used containerization projects such as Docker and Kubernetes are popular amongst developers. It has a vibrant user and contributor base and it is driving service portability across platforms.
In addition, open source software-defined storage solutions such as CEPH, along with streaming analytics platforms like Kafka and Spark, and even AI frameworks like TensorFlow, will help to drive the development of applications and platforms in edge computing.
Open source will drive interoperability by its very nature. After all, no developer would want to build dozens of different versions of the same app to suit different operators and enterprises – over different platforms. Operators must adopt common IaaS, Platform as a Service (PaaS), and orchestration architectures. This strategy will ultimately lower costs and drive focused innovation. What’s more, standardization of platforms across operators and enterprises will be more formally driven by CNCF (Cloud Native Computing Foundation) projects such as Prometheus, Envoy and Linkerd.
The problems with projects
Nothing’s ever entirely straightforward, however with the edge. There are issues of scale and complexity to be resolved and virtual NFV/SDN and telco-centric orchestration systems won’t be entirely familiar to edge application developers accustomed to the cloud. As some developers will not necessarily understand the complexity of a carrier network, Linux Foundation’s ONAP (open network automation platform) has considered the need for providing a lightweight, developer-centric orchestration and marketplace platform.
Ultimately, open source projects have always been a ripe source of innovation. New domains however can bring new and often competing projects which can lead to the creation of silos. With a need for communities to work more closely together, consolidation is therefore a critical area for future development. Communities can also lose focus when too many projects are in development at the same time. How can you avoid this? Focus on the three pillars of data center technologies: compute, network, and storage. By concentrating on the needs of developers and removing as much friction as possible from the process – open source can ease the transition to the edge. It can benefit operators and their subscribers – and enterprises and their users.
