Poor cloud strategy can lead to unintended silos

Poor cloud strategy can lead to unintended silos

The more things change, the more they stay the same.

In the last few years, IT has undergone a massive shift as hybrid cloud and multi-cloud have become entrenched in most organizations. By housing application workloads across a mix of public and private cloud platforms, with containerized orchestration among the environments, companies can maximize cost efficiency, scalability, performance, and other benefits.

However, the growing number of disparate clouds is raising a concern that has dogged IT forever: silos.

In the traditional on-premise world, silos proliferated when enterprise data and applications were spread among fragmented servers and data centers that were unable to easily interact with each other. The hodgepodge was a nightmare to maintain, diminished efficiency and balkanized data rather than making it a company-wide resource.

While the cloud is a game-changer in how organizations manage their IT infrastructures, it is prone to silos as well. In fact, perhaps even more so. Simply by breaking out a credit card and clicking a button, individual business units or departments can implement a business application on one cloud or another, without coordinating with the enterprise as a whole.

It’s a pretty simple formula: The more clouds you have, the more control planes you must manage. Complexity increases.

This cloud sprawl likely will intensify as pressure mounts on the major cloud vendors to avoid commoditization and begin to compete with new services around workload migration, AI, and other capabilities. This increased differentiation will only heighten the appeal of different clouds to different parts of the company, as well as embolden habitual “collectors” – i.e. a decision-maker loyal to AWS in one business unit while someone in another always buys from Google.

Organizations need to thwart the proliferation of cloud silos, as it will prevent them from fulfilling the reasons they adopted hybrid and multi-cloud in the first place and hinder their ability to take advantage of the cloud for compute-intensive workloads like AI and data analytics.

In the past, organizations dealt with silos through heavily customized integration software – a slow and labor-intensive process. That approach is a non-starter in cloud computing, which is supposed to be all about efficiency, agility, and turnkey.

It’s crucial for every organization today to have a plan to operate in a multi-cloud way, and to minimize silos. That requires determining and implementing the right level of abstraction for application orchestration against multiple cloud endpoints.

Technology exists that is squarely aimed at solving the problem of multiple substrates. Users can choose where to deploy a model of an application, and continue to manage their application stacks in a predictable and identical fashion, independent of the cloud substrate they are running on.

Freed from the complexity of the different control planes, management and administration of many workloads across clouds become drastically simpler. Enterprises can leverage the differentiators of each cloud provider without sacrificing operational predictability.

This approach ameliorates another issue with cloud silos – the skills challenge that comes with trying to manage multiple clouds and associated technologies like OpenStack and Kubernetes.

As multi-cloud becomes the reality in today’s enterprises, it’s important that organizations steer clear of the mistakes of the past and minimize silos. That’s the only way to fully reap cloud’s rewards in increasing efficiency, fostering innovation and meeting customers’ needs.

Read next: The UK government's settled status app for EU nationals is an embarrassment