Avoiding the Risk of a “One-Size-Fits-All” Approach  to Cloud Migration

Many organizations are working hard to move to the cloud, but find that with a migration there is also complexity. Recently, Derek Swanson, CTO of Silk, offered advice on what to evaluate to successfully take advantage of all cloud has to offer, the issues to consider when determining what infrastructure will best serve each workload, and the risks of going to the cloud with the wrong strategy.

Swanson has more than 25 years of experience as a technology evangelist, systems architect, and data systems engineer. At Silk, he is responsible for guiding the customer-facing architect teams, developing the product road map, and is the primary technical evangelist in the organization.

What does Silk provide? What is its unique value proposition?

Derek SwansonThe Silk Cloud Platform is a virtualization layer that fits between your mission-critical databases and applications and the underlying cloud infrastructure, delivering up to 10x faster application performance compared to native cloud alone. Silk also makes it possible to accelerate cloud adoption timelines of these mission-critical workloads by enabling a simple lift and shift of the entire workload onto the cloud as-is with no need to refactor.  The unique value proposition is that Silk provides an enterprise-class cloud data platform that provides the performance, resilience, availability, data services, and vertical scalability that big mission-critical databases need. These features are not natively available from the public clouds, so Silk enables companies to adopt the cloud immediately and achieve the benefits without having to refactor or modernize their application stack before they can move.

What is the difference between a cloud-smart and cloud-first approach?

A cloud-first approach means that we want to build and run all applications on the cloud, regardless of their function or use case. Cloud-first approaches will often run into cost or performance challenges as some workloads are better suited for on-prem architectural capabilities and don’t fit well into the cloud without deploying inefficient architectures.

A cloud-smart strategy is a far more flexible approach that allows companies with varying business needs to move applications on a case basis from on-prem to the cloud (or vice versa). Application workloads are analyzed beforehand to see how well they will potentially fit into cloud services. Cloud-smart enables a cloud strategy that aligns with specific business goals and values while maximizing cost efficiency and minimizing risk exposure. According to Gartner, 60% of organizations that have adopted a cloud-first strategy will replace it with a cloud-smart strategy by 2023.

What are some of the mistakes you see companies making in their transition to the cloud?

Cloud spending is on track to surpass $1 trillion by 2024, partly due to urgent changes to business operations driven by the pandemic. CIOs and IT leaders are rushing to advance cloud strategies, but implementations often may not be fully successful given the deep complexity of many mission-critical applications. Cloud migration is not a “one-size-fits-all” job, and the public clouds are also not designed to fit everything easily into their frameworks. Though migrating some workloads does prove more straightforward than others, moving mission-critical data can introduce considerable challenges if their complexity isn’t properly understood, and the high visibility that accompanies these major initiatives dramatically increases the consequences of failure. Mission-critical applications are essential for business value, and steady, reliable high performance is a must-have for operations. Mission-critical applications cannot become unavailable either due to “slowness” or an outage.  The risks and complexities around rushing through a mission-critical migration project are significant.

What is causing this?

The pandemic drove a couple of major shifts in public and private behavior.  There was a mass exodus from offices to remote working, shifting the traditional office environment to a distributed workforce. Next, people decided to self-isolate, staying at home and transitioning many tasks to remote help—doing a lot more stuff online and going out a lot less.  The massively increased volumes of both remote work and remote day-to-day life by using online resources to deliver things like food, healthcare, and entertainment led to a surge in need for cloud-based applications and databases in an extremely compressed time frame.

What does the C-suite expect and what is really happening?

The C-suite perhaps thinks that migrating all workloads to the cloud can happen fairly easily and fairly quickly because they don’t understand how nuanced the process is and that this isn’t a one-size-fits-all situation. While some workloads can simply be lifted and shifted to the cloud, others that are larger, more complex, and oftentimes, mission-critical are just too extensive to get there through a simple lift-and-shift—or frankly, at all. Refactoring or modernizing simple applications while rationalizing less useful apps is usually a great strategy to begin to move operations to the cloud quickly. 

Challenges arise when the company needs to move a big “anchor” workload that has dozens or even hundreds of other applications dependent upon it.  These anchor workloads have extremely high “data gravity”, which means that they are too heavy and resource-demanding to run on cloud-native solutions (either IaaS or PaaS) and therefore cannot move. They are preventing a great number of smaller applications that could run on the cloud from moving because those applications are dependent on the anchor workload being physically close enough (with low enough latency and/or high enough throughput) to run properly. The reality is that often times organizations end up a bit stuck in the middle, where they are unable to move everything to the cloud they would like, and they cannot move back out of the cloud—their cloud-first strategy is stalled out, and they end up running in a hybrid configuration, partly in and partly out of the cloud, paying for both on-prem and cloud commitments.  It is a really frustrating place to land.

What is the risk of moving to the cloud with the wrong strategy?

If you use the wrong strategy to move workloads to the cloud, the consequences can be dire. Best case failure scenario: your migration plan gets stalled because you can’t simply lift and shift a workload onto the cloud. So you have to switch your strategy from lift and shift to a full, time-consuming refactor.

Worst case scenario: You manage to get your workload into the cloud and find that you don’t get the performance or feature/functionality that you and your end-users need. The slower performance or lack of features can literally shutter your business if the degradation in customer experience becomes significant enough that people stop using your services.  This happens quite frequently, especially with traditional companies that have a large suite of line-of-business apps they are trying to move, competing with new startups that are building greenfield applications using a cloud-first strategy without dragging along any legacy technical debt.

What are the advantages for companies that do it right?

For companies who do it right, the benefits of the cloud are theirs for the taking: scalability, agility, expanded reach, simpler 24/7 operations, better value from operating costs, and improved DevOps cycles for delivering new products and improved customer experiences.

What should organizations consider for long-term cloud success?

For long-term cloud success, consider each application individually and whether it can or should go to the cloud or if it’s better suited for on-prem or perhaps rationalization. Legacy applications that are looking at retirement in the next few years might not be as good a candidate for the cloud, as the migration process will be more trouble than it’s worth. For large, complex workloads that leverage big vertical stacks with Oracle and Microsoft SQL, consider moving to the cloud using a cloud data platform that offers faster performance, availability, and high-value data services than is available from the native cloud alone.


Subscribe to Big Data Quarterly E-Edition