Serverless and Multi-Cloud

The Multi-Cloud is fast gaining traction among the enterprises and the analyst firm Gartner has been talking about how different organizations are embracing more than one public cloud provider. Serverless is fast gaining traction among the developers as the right abstraction. The abstraction offered by Serverless and some of the restrictions imposed on the applications makes Serverless less attractive in the Multi-Cloud world. In this blog post we will try to address this question and talk about how Serverless can make sense in such deployment models.

Why Serverless doesn’t make sense in Multi-Cloud?

The Serverless abstraction is all about reducing the operations burden on the developer and let them focus only on codifying the business logic. Developers are not expected to care about what is under the hood of a Serverless platform but Multi-Cloud imposes additional operational overhead to developers. This overhead neutralizes any productivity or agility advantage developers get from Serverless platforms. If the underlying driver is about removing operational overhead and give developers a seamless way to deploy their applications, a multi-cloud environment is not a good idea.

Serverless in Multi-Cloud: Does it make any sense?

Having pointed out how Serverless makes little sense in the context of Multi-Cloud, we need to highlight the nuance that comes in the context of using multiple cloud providers. This is especially true for enterprise customers where Multi-Cloud is the foundation of their IT environments. Now the key consideration is how they can provide a standardized environment with the Serverless abstraction in order to improve the productivity of their developers and ensure an agile culture. In this case, Serverless offerings provided by Hyperscalers make limited sense. It may fit for specific applications used in the organization but they cannot work across multiple cloud providers.

In order to provide a Serverless like abstraction across multiple providers, organizations will have to look into open source ecosystem. With Kubernetes gaining significant traction in the enterprise, a Serverless abstraction on top of Kubernetes seems like an obvious choice. Knative is an open source project that provides this abstraction on top of Kubernetes. Knative just reached the Version 1.0 milestone and we can expect more organizations to embrace the platform. However, Knative is still not as mature as OpenWhisk which has been there for a longer period and offers support for more tools needed for applications. OpenFaaS, Kubeless, Fn, etc. are other open source platforms that are useful in the Multi-Cloud context.

The main driver for Multi-Cloud Serverless platforms is the need to standardize the developer experience across team. This standardization removes friction, reduces costs in training talent and offers other advantages for large enterprises. While the operations overhead is a cost incurred by organizations deploying Serverless Platforms across Multi-Cloud, the other benefits give them an advantage over container based on VM based platforms.

Share this post

Share your thoughts!!

This site uses Akismet to reduce spam. Learn how your comment data is processed.