Chicken or Egg!
At the centre of discussions at last week’s Edge Congress in London was the question of what needed to come first in making the CSP business case for edge cloud: Compelling edge applications or available edge infrastructure. The presentations largely focused on compelling applications for edge cloud, but in an interesting ad-hoc panel session the majority of panel members said that edge infrastructure was key to driving demand.
Appledore believe that the underlying business case for deploying edge cloud will fundamentally be based on infrastructure economics. Edge Cloud needs to address the same basic infrastructure issues as public cloud:
- Maximize multi-tenancy
- Minimize power and environment infrastructure and running cost
- Maximize automation and minimise operational cost.
Achieving this cannot rely on one or two “killer applications”. In fact a reliance on one or two applications is highly likely to drive the edge cloud to specifically support these applications. At the same time, without a compelling base application, it is unlikely that CSPs or anyone will invest in edge infrastructure.
Chicken and Egg!
Squaring this circle the CSPs can learn from the public cloud players. AWS infrastructure was not created based on speculative new applications; rather it was the repurposing of infrastructure from an existing eCommerce store. Amazon had both the infrastructure and the base application. Innovative new applications, like Netflix, came about because of that existing infrastructure, and the building of a virtuous circle of applications and infrastructure investment.
CSP actually have a similar opportunity with the real and immediate need for telco edge, requiring deployment of edge infrastructure at central offices and base station sites. The challenge is that this opportunity will be lost if we continue the boundary between what is a network application and function and what is a software application and function. Segmenting network applications from any other application inherently means that you lose the opportunity to maximise multi-tenancy and minimize power and environment cost. It also means that you duplicate operational resources and entrench existing manual network management processes rather than efficient cloud native approaches.
CSPs have a limited window of opportunity to win in public edge cloud, but only if they can overcome the cultural boundary between what is network and what is everything else. CSPs need to invest in an edge infrastructure capable of managing any application based on their core application the network. In talking with operators at the show, some are already recognising this and making this change.
We explore the potential for the optimisation of edge infrastructure with access connectivity in our recent research report Edge Cloud- a new CSP business case and will be publishing further on the business case for edge later in 2019. We report on the key challenges in achieving cloud native in CSPs in Cloud native – a revolution postponed.