The Multicloud approach

Over the past couple of years, we are increasingly seeing our customers and partners adopting the multi-cloud approach. Depending on the profile of the business and nature of the program, here are some of the key drivers (or at times a fait-accompli) we have come across:

1. Vendor flexibility ie. no vendor lock-in

Admittedly, it ain’t as big a deal as it used to be in the on-premise world, but it still figures high on the list of CIOs - especially in big enterprises. Concerns over continuity in pricing/pricing model, SLAs, rate of innovation et al. figure prominently.

2. Choice of service/solution

“AI suite of GCP wins hands down but we still find the app platform services of AWS to be a better fit.” - we hear similar requirements all the time. For instance, a customer building an Augmented EHR with us wanted to leverage GCP’s Healthcare APIs for FHIR / HL7 but had invested significantly in building their ML / MLOps stack on AWS. For a variety of reasons and factors, the ability to cherry-pick services and solutions from different vendors is desirable at the very least.

3. Mergers' &Acquisitions'

At least four of our customers have had to adopt and embrace the multi-cloud approach because of the acquisitions they made. We have witnessed a drastic reduction in ops and engineering teams’ productivity in at least two quarters following the integration of systems, tools, and processes.

4. Enabling wider adoption

Enterprise SaaS and enterprise cloud software in general which often gets deployed in the customer clouds need to have the flexibility to get deployed across various clouds and private cloud stacks. Provider’s ability to deploy, monitor, maintain, and upgrade is crucial to enable wider customer adoption.

So, what are the different aspects one must consider to be in the state of reasonable readiness to adopt a multi-cloud approach:

1. If you are an enterprise with a buy > build approach, multi-cloud support in the following key solutions is a must:

a. Infrastructure monitoring

b. Backup and DR

c. Data governance

d. Data warehousing / Data lakes

e. Compliance and regulatory software

f. MLOps

g. Identity and access management

2. If you are building a SaaS / Cloud-first service, explore:

a. Multi-cloud CI/CD => think Hashicorp Terraform?

b. Containerization and/or Kubernetes-based underlying infrastructure

c. Right abstractions for data pipelining - cloud agnostic orchestrations - Airflow as an example - should be preferred over cloud provider-specific solution wherever possible

d. Data access layers with abstractions

Are you ready to future-proof your organization with a Multi-Cloud solution? Reach out to us on [email protected] to know more.