In part one, we looked at the different types of clouds in 2024, and what Supercloud is and is not in this ever-growing field. The benefits it promises are huge, and many organisations would snap them up in seconds. The question we posed at the end of part 1 was: why is it not the standard for every organisation? The simple answer: it is not easy!Â
Let’s look at some of the issues customers face on this journey.Â
The complexity IssueÂ
Creating a true Supercloud is hard work and takes a lot of dedicated engineering time, but the results can impressive One of the core complexities for organisations adopting Supercloud in 2024 will be existing technical debt. We are unlikely to get the option for a greenfield deployment, meaning a lot of organic growth and a decade of cloud technical debt will need to be unravelled. Â
Additionally, each cloud comes with its own set of tools, APIs, and management interfaces. This moves a lot of requirements to an orchestration tool that can interface with each, but also puts the work on the technical teams to code the outcomes. This is the time and cost part, which means organisations must make a conscious decision before starting the journey. If you are struggling with one cloud, maybe leave Supercloud for a while!Â
The biggest challengesÂ
To extract the value from Supercloud, you need to either be able to run workloads across clouds (scaling up or down at the most appropriate) or move workloads between the clouds (to take advantage of pricing or features). The technical capability to move workloads from one cloud platform to another, for many IT admins, would be a scary thought. Especially for those still in the world of virtual machines (more on that later).Â
There are a few other considerations before embarking on this type of architecture decision; it is about more than just the technical ability. We need to consider all the associated business processes that would have to move in parallel to the workloads and the impact that could have on service delivery. Â
Data protection and compliance:Â
For many organisations, data sovereignty can be a significant challenge; adding Supercloud's flexibility could add additional headaches for compliance and risk teams. Also, considerations need to be made around data protection and retention policies. As applications move between platforms, who is responsible for the backup and retention of that data? Compliance with your data policy and regulations like GDPR must be factored in.Â
Accounting and finance processes:Â
Changing cloud suppliers will impact finance processes, procurement, and spending commitment. Payment processes, sign-off, and technology will dictate the ability of an organisation to switch cloud suppliers quickly. Monitoring and understanding cloud offerings can be difficult, as pricing and features change daily, and many services and offerings have no direct equivalent in competing clouds, making the process harder.Â
Many cloud providers only provide the best pricing on long-term subscription options, which is contrary to the concept of workload mobility (though it does open the idea of spot instances that can offer significant commercial advantages)Â
Software licensing:Â Â
Looking at the entire software ecosystem can often be very complex for many organisations; adding portability to multiple cloud providers will magnify the considerations. Most providers don’t want to make it easy to move to the competition.Â
Skills and operational processes:Â
Each cloud requires new skills, different processes and workflows, and the possible complexity of new security products and services. While the Supercloud architecture will help users consume services more efficiently and organisations take advantage of new capabilities, a technical resource still needs to code the outcome!Â
Performance and latency Issues:Â
Building resilient, performant, secure applications that are also easy to manage based on one cloud platform can be a full-time job. Bringing in the challenges of cross-cloud network latencies, and different underlying compute and storage performance characteristics will need to be added to the mix.Â
Consistent quality of service:Â
One final thing: someone is responsible for the SLA of the application services your co-workers and customers consume. Moving services between clouds could add risk to those SLAs, and of course, it could also increase availability if it is architected well!Â
Portability of workloads:Â
This is the big one regarding making Supercloud work or not: the type of applications and services you are running. Â
If you are mainly developing or running applications inside virtual machines, I believe Supercloud (in fact, the same goes for multi- and hybrid clouds) is not a good fit. Virtual machines are from an era that had not dreamt of concepts like Supercloud, and as such, they bring a wealth of complexities that I think most system admins would agree are not worth the risk, time, or cost. Â
Now, container-based, serverless and microservices-based applications are much better-placed to work within Supercloud architectures. Supercloud has become more relevant in recent years due to the increase in modern application architectures. Gartner tells us, "By 2027, more than 90% of global organisations will be running containerised applications in production, which is a significant increase from fewer than 40% in 2021" This leap in adoption makes Supercloud significantly more relevant than just a couple of years ago.Â
SummaryÂ
Sounds like hard work! I usually recommend that most organisations adopt a single public cloud policy (when regulation allows) and keep life simple. The Supercloud benefits require scale to become viable and, for most organisations, need more practical achievement. Hybrid Cloud, for sure, is worth significant consideration. It can be much easier to adopt and integrate; also, when we consider new edge use requirements and the future of AI Inferencing, I would expect many organisations to end up in a true Hybrid Cloud architecture (remember, a consistent operating model is needed to make it genuinely hybrid).Â
Now we understand what it is and the challenges it gives us a platform to make informed decisions on adopting a viable Supercloud Architecture; join us in Part 3, and we will explore how we can go on this journey. For those who don’t need it, some of the concepts for Supercloud are the same for hybrid, so read on as well.Â
Contributors
-
Rob Sims
Chief Technologist - Hybrid Platforms