Understanding of Edge Computing Models and Strategic Considerations

Cloud computing has been beneficial for every organization who successfully adopted it. Different organization used the on-demand services and models (SaaS, PaaS and IaaS) offered by cloud for handling different operational processes. Cloud adoption was accelerated in the Pandemic. But along with it, a new focus area has been emerged prominently in the last 3-4 years in form of edge computing.

By now, we know that edge is any mini server processing data locally to support the new age technology use cases like IoT, autonomous cars, AR/VR, smart city, etc that require real-time processing of information. Due to more response time occurring with geographical distance, Cloud computing architecture was not the right choice for new-age use cases. New-age technology use cases demand high bandwidth and low latency that is ultimately provided with edge computing.

Understanding the edge

In 2022, the edge computing concept is known to the CIOs and leaders of many companies, and everyone has their own understanding. For many companies, edge computing is the natural extension of cloud computing and can be considered as a strategic move in having a hybrid cloud architecture. Another perspective is that an edge and cloud work hand in hand to provide on-demand services from everyone in the world as well as closer to devices.

In a recent presentation by Jim Morrish on edge computing, the definitions of edge computing are nicely explained. Edge computing is different for different industries, but the common theme is to move processing/computing closer to where devices or users are that require instant and real-time intelligence information.

Like definitions, new terminologies are also tossed up. For telecom operators, the term Mobile Edge Computing (MEC) is evolved into Multi-access edge computing.

In my opinion, I see edge computing is gaining good attention from organizations who rely on their own data centers and some part in the public cloud, and also those who are looking to evolve their IT infrastructure to tap new business opportunities. Apart from this, a few telecom operators like Rakuten, Verizon, SK Telecom, China Mobile, Duetsche Telecom, etc have started providing edge services in their network and several are collaborating with solutions vendors to build their edge stack. In the next few years, in providing the supreme 5G services to end users, we will see edge deployments from telecom operators, starting with private 5G offerings in some countries.

?Key things to consider with Edge deployment

Planning the edge deployments sounds simple and implementing it is not straightforward. This is big call if you are going to implement edge as you invest your time, money and efforts. As we are in transition, we need to consider some of the key implications of deploying the edge in any IT infrastructure.

Automation and remote management are very crucial

Edge nodes can be any number of locations depending on who is incorporating it. This is a mini computing infrastructure that you cannot always deploy IT staff to manually configure, upgrade resources, apply policies, and scale. Automation can help by setting up a centralized system where you can define repeatable processes and instructions to reduce the need for manual operations by IT staff. It is helpful in improving provision of IT resources, automatically perform patching, security configurations and keeping edge server environments up to date. In case of edge, the automation workflows can be set at the edge endpoints and centralized layer. Such localization can be helpful in mitigating several issues at the edge when manual debugging is time consuming.

Automation also needs to be supported with proper management of resources to standardise the edge environments. If we take an example of the retail store where edge nodes are deployed, defining and automating a standard image across branches is all you need to do. This way we can perform configuration on multiple branches, better plan for disaster recovery and responding to certain events with automation.

No Edge with Kubernetes

When we think of an edge, consistency and remote lifecycle management of your software applications deployed at the edge become important. You want consistency from centralized cloud to all the participating edge nodes. For example, ML modes trained at the central cloud push to the edge to eliminate the need for traversing computing requests in large networks and that is the clear purpose of deploying the edge. But in this case, you need to bring the consistency in pushing right set of ML workloads and lifecycle management of all ML applications along with rest of software applications that will process the data locally. Kubernetes can be helpful in this case to host the lightweight containerised applications and perform all required things to improve software lifecycle management along with enabling consistency. Not just containers, Kubernetes can be used to orchestrate the applications deployed in virtual machines.

Cannot ignore Observability and Security

?The next thing after automation and remote management to focus is observability, monitoring and security. Edge deployments creates more points of failure and potential for external security threats with its scale in numbers. The primary purpose of edge deployments to provide performance to company objectives by processing critical data close to devices. By gaining visibility into edge service environment, you need to ensure that applications are meeting goals defined during implementation of edge. Also, it is important to ensure that any malicious intrusion and access need to be hunt down before it can make a damage to data.

#edge #edgecomputing #kubernetes #edgecloud #cloudcomputing #distributedcloud #5G #technology #CIO #CCO