Join us
Login

Avoiding cloud lock in requires a pragmatic approach: Kubernetes may not be a panacea

Published by The Stack
July 21, 2021 @ 9:00 AM

The Stack is one of the UK’s fastest growing business technology publications. Founded in late 2020 by Edward Targett, the former editor of Computer Business Review, it regularly features interviews with some of the world’s most influential digital leaders, along with think pieces, news, and business intelligence. You can follow The Stack on LinkedIn here.

____________

 
Avoiding cloud lock in requires a pragmatic approach

Building your own cloud infrastructure after years spent investing in a vendor’s platform takes a monumental effort. Despite mounting costs, it’s understandable why firms often spend decades locked into services that no longer serve their business needs, writes Tom Christensen, Global Technology Advisor and Executive Analyst, Hitachi Vantara. Services like Kubernetes have long been touted as the answer to avoiding vendor lock-in but require a change in application architecture towards a modern app architecture. Kubernetes could prove disastrous if not approached with clear strategy and direction. To avoid vendor lock-in whilst simultaneously avoiding siloes between application developers, operation, and security team, follow these steps: 

Never rush to cloud  

Over the past year, a significant level of attention has been placed on the rapid digital transformation efforts of enterprises around the world as they adjust to the restrictions the pandemic has placed on us. You might have seen that as part of these efforts, many businesses have been quick to invest in cloud infrastructure, perhaps even for the first time. Looking beyond the hype that has and continues to surround it, rushing to cloud – even during times of crisis – isn’t always the best decision for every business, and could result in a disastrous and costly fate.   

For example, not long ago, I worked with a big system integrator in the Nordic region that pledged to move 80% of its workloads into the public cloud. After three years of arduous attempts, they only managed to move 10% before abandoning the entire transition project and stick to on-premise. This is just one example of the many cases where companies will recognise far too late that the move to public cloud is a costly endeavour, leading many to pivot back to their data centre roots.  

To put things into perspective, it’s also worth remembering that a number of financial services organisations have estimated that it will take an average of 35 years to modernise their applications at the rate of current transformation. Whilst many businesses may feel like they’re playing catch-up, they ultimately shouldn’t be pressured into rushing the modernisation process. 

Prioritise and strategise  

Beyond the burden of costs, the example also exposes a tendency for businesses to take an all-or-nothing approach to cloud investment. To avoid disappointment, businesses should instead be looking to first uncover where the biggest transformational benefits lie, channelling their investment in areas that matter most. I would always advise others to first prioritise the applications that return most value to a business and can provide the biggest return on investment. As a general rule of thumb, businesses should be looking for the areas that can guarantee: 

·       Fast delivery of modern applications 

·       A robust platform to host modern applications 

·       Quick deployment  

·       An ability to meet customer and business expectations 

·       Business agility 

Priorities will of course vary from business to business but, generally speaking, it’ll be the businesses that can recognise and understand which areas of their business to prioritise their investment that will go on to form the most robust and risk-averse cloud strategy.  

Avoiding cloud lock in: Know when to say no 

Likewise, by enabling users to securely manage and monitor their data across major cloud providers and on premises, it’s unsurprising that Kubernetes is often considered a go-to choice for distributed applications. However, when managing legacy static and mainstream applications, the approach is often time consuming and heavy-handed. As a platform designed to improve performance, it’s wise to be selective when choosing applications for Kubernetes modernisation.  

For example, I have worked with a number of technology leaders looking for cloud-based management tools which come with the flexibility to transition to a multicloud or consumption-based model. In cases like these, there’s a need to manage containers at scale, and without leaving current levels of investment in IT behind. In these circumstances, Kubernetes would be the smart choice, and I would advise leaders to focus on supporting DevOps and DevSecOps and establishing a hybrid-cloud platform for running their microservices. 

Simplicity is key  

Avoiding cloud lock-in is a challenging thing to do, so the best advice I can provide is to keep things simple wherever possible. Having spoken to a range of IT businesses and leaders over the past year, it’s clear that many businesses’ concerns regarding Kubernetes deployment surround the complexity, security, management, and automation of an agnostic Kubernetes service. It is for this reason that simplifying the deployment of the platform and closing skill gaps where possible will be your key to success when it comes to avoiding vendor lock-in.  

Overall, cloud expertise will be your best friend during this process, along with an understanding that every business and its varying needs requires a tailored and measured approach. Whether it involves your team undergoing certification-level training, or a company-wide recognition of the business verticals that matter most, these steps could prove instrumental in the success of your deployment and will play an important role in driving the business results you want. 

See also: AWS turns from open source villain to hero (at least, for some Elasticsearch users)…

Posted in:CIONET UK

Subscribe by Email

No Comments Yet

Let us know what you think