By Andrew Weaver, Lead Specialist Solutions Architect, Databricks
The past two years have shed light on just how critical a digital-first cloud strategy is in keeping a distributed workforce up and running. According to a recent Gartner study of public cloud users, 81% of respondents said they are working with two or more cloud providers, which just shows how prolific multi-cloud has become to business continuity and growth today. There is also a new breed of technology companies emerging that need to sit across all three major clouds to be truly successful – making the multi-cloud approach non-negotiable. And at the same time, cloud vendors all want to create the best cloud environment, where apps can sit seamlessly across the top and customers can keep their data where it is to perform Artificial Intelligence (AI) and Machine Learning (ML). So, what does this all mean for multi-cloud?
I predict that, over time, we will move towards ‘true multi-cloud’. True multi-cloud is, simply, when the end user can safely and without duplication do any job, on any data, regardless of the format, without even having to know which underlying cloud provider they’re using at any given moment. There will be no need to navigate any complexities, such as the networking or IAM model – they will simply be able to go about their working day as the cloud beneath them continues to switch and change, seamlessly.
But how far away are we actually from this ideal? And, if it does come about, how will organisations best be able to tap into its advantages?
The ease of true multi-cloud
Let’s start with how true multi-cloud can benefit businesses and the teams within them. Firstly, organisations’ skilled data scientists, engineers, analysts and business users will spend less time having to administer the different clouds, move data around, set up network and IAM configurations, and more time focusing on providing value to the business. Secondly, true multi-cloud could result in a significantly reduced tech footprint for organisations, with a reliance on tools that are universal as opposed to specific to each cloud. This will likely have a cost benefit but it could also mean there is a reduced need for training – both on the tools themselves and the underlying cloud provider that they run on. Employees will be able to get up to speed in less time and with fewer resources, enabling a more fluid, uninterrupted working day, as well as allowing far more time to be set aside for business activities – such as key innovation and the decision making that this drives.
For organisations seeking to lay the groundwork and tap into this to the fullest extent, moving away from data warehouses and adopting a data lakehouse architecture can offer great advantages. A data lakehouse provides easier management, greater governance and efficiency for organisations operating on more than one cloud. As true multi-cloud comes into its own, organisations using a data lakehouse will be able to work with their data without having to continually master the complexities of the underlying cloud. With a lakehouse, organisations can build pipelines and access patterns to collect and enrich data, generate and share insights and even train ML models in an agnostic way that will run on any given cloud.
What true multi-cloud will look like for cloud vendors
Returning to cost, this is another area where organisations could see significant benefits if true multi-cloud were to take shape. None of the clouds are going to ‘win out’ and replace the other ones – they are all here to stay, even if they will always compete. But in this true multi-cloud future, they will likely compete primarily on features and costs, rather than locking in all of the data and workloads for a given customer. For organisations operating across different clouds, this could see a net benefit of better services at a more competitive price. There has already been some movement in this direction. At AWS re:Invent 2021, for instance, the price reduction for data transfers out to the internet was announced – signalling the lowering of egress fees.
All this cuts away at complexity, giving businesses greater access and control over their data. This could ultimately benefit the cloud providers too as they are trying to win the ‘data storage’ race, particularly as more data & AI companies emerge. Cloud vendors want customers to have seamless cloud ecosystems, where they can easily access data and leverage AI and ML services wherever their data resides, and true multi-cloud offers a way of making this a reality.
To be clear, this vision of true multi-cloud in its purest form is not something that will emerge overnight – to achieve it cloud providers will need to be able to orchestrate and leverage their services on another cloud. It would be far easier for a cloud neutral platform to provide the same effect, by interconnecting all the various platforms, tools and clouds and providing clear visibility and governance across it all. There is still certainly a long way to go, but with the benefits it can offer to both customers and cloud providers alike, true multi-cloud may be arriving on our doorsteps sooner than we think. For organisations seeking to prepare, looking inwards at their data architecture and management is a key place to start.