Accounting for technical challenges in edge cloud development is critical to understanding the evolution of the edge cloud over time. This area remains ill defined today as confusion reigns in the market. Part of the reason comes from the large number of applications that benefit from edge computing. These applications come from different industry and consumer verticals, leading to different types of edge cloud implementations. Players from different backgrounds that include cloud players, telcos, and enterprises are in process of developing these approaches, which hopefully would converge at some point in the future.
The Technical Challenges of Edge Cloud
Considering the edge cloud as decentralization of computing, storage and networking resources, we can identify the following technical challenges:
- Orchestration of services over multiple sites. One of the advantages of the cloud is that the user does not have to worry about where the application is running. The hyper-scalars operate a few data centers with massive scale – applications would be running in few data centers. In the case of the edge, applications may be running of many (hundreds, thousands?) of locations. This results in a large number of control planes. Stitching these control planes for seamless operation of applications is a challenge.
- Support for multiple types of hardware. Data centers servers rely on x86 processors. On the other hand, the edge cloud includes different types of hardware, such as GPUs or FPGAs for example. It becomes critical for the edge cloud stack to support different type of hardware on demand per the application requirements.
- Remote management, automation and zero-touch provisioning. The edge cloud requires installing and upgrading software remotely without the presence of skilled staff on site, in addition to the ability to perform these tasks on multiple sites that could range in the thousands which is very critical.
- High software stability and availability. The cloud benefits from the high scale of data center to provide high availability and stability. The small size of the edge heightens the importance of software stability and availability which needs to be higher than that of cloud data centers.
- Security. Security cuts across different functions which includes, for example, secure communications, security of data at rest and in motion, trust between devices and the data center, and identifying and stoping rogue connections. Moreover, physical security is very critical since edge data centers could be located ‘anywhere’!
- Support for multi-tenancy at the edge. Similar to the cloud, the edge needs to support multi-tenancy.
- Thin control plane and stack. The edge is ‘small’ – space, power, compute and storage are limited. The software stack needs to be efficient and scalable to meet the operating characteristics of the edge.
The Business Aspects
A number of companies and open source projects are working on resolving the technical challenges in edge cloud development listed above. Their commercial success will determine how the edge cloud will evolve. The ones that success would have balanced the benefits of edge cloud with the cost of ownership [see here and here]. Cost has to factor the physical location, power and communications requirement among other things. Here, business questions come strongly into the picture: how to deal with a heterogeneity of edge cloud platforms, for instance? This is a topic for another day!
In part of assessing the evolution of edge computing, Xona Partners has developed a set of frameworks based on technical, competitive and financial dimensions to gauge ecosystem readiness and market viability for corporate and financial investors. Contact me if you are interested in our service in this area.