Article: The Cloud’s Tough Pill to Swallow
by Chris Swafford
Chris Swafford is a Consulting Architect at Fulcrum Technology Solutions. Swafford has been a part of the Infrastructure team since 2014.
It is 2019 and almost everyone has seen the advertisements: the “Cloud” is here for you and your business, and it will allow you to be fast, nimble, and secure. It is the source of all potential for your company’s future growth. If the “Cloud” were a super hero, it would be Superman. And sure, there is indeed great potential with Cloud, and a company’s IT department would be short-sighted if it were not considering how to leverage this potential. But as with all new technology, there are tough pills that you must swallow to be successful with it.
All workloads are created equal, but some are more equal than others
Marketing departments for cloud providers are very good with their messaging. I have heard too many times that a company wants to “lift and shift” their datacenter into Cloud because they saw the advertisements. When pressed for details about what results they are looking for and what requirements they have, they often cannot provide them — they are convinced Cloud is the future and they want to be a part of it.
If everything you run is based on standard Intel processors and chipsets, then sure, you may be able to put the workload in a cloud provider’s datacenter. But what about that 2-tier finance package? Or that one last Oracle Database residing on a Sparc system? Some providers can run Sparc emulators, but do you really want that for your critical Oracle database? So perhaps you decide to leave those exceptions in your datacenter… but now you have introduced latency in each database call – because maybe you moved a front-end workload while leaving your Oracle database on-premise. You can only communicate at the ‘speed of light’ within a datacenter: within your own datacenter or between your systems running in a Cloud provider. Latency-sensitive systems like front-end and back-end systems will suffer if you place components in different datacenters. And similarly, if you have latency-sensitive systems like a 2-tier ERP, your users who are running a thick client on their workstations will now complain about how slow the application is since you moved to the Cloud because their back-end system is now located in a different datacenter.
That potential success you envisioned has now been lost because not enough planning was completed.
Your success is my success, your sorrow is my sorrow
I have heard the sales pitch in board rooms many times: their cloud solution is growing by leaps and bounds and new services are coming online monthly. “Join us,” they say, “and our success will be your success.”
That sounds great! Who doesn’t want success? But fast growth does not come without some pain. Recently, an entire Microsoft Azure regional datacenter went temporarily offline due to a lightning strike. Azure does not have magical datacenters and they are just as susceptible to catastrophic failures as your own datacenter might be. Sure, their redundancy is top of the line, but you should not blindly put your trust in their designs. As the customer, you must design for potential outages. There will be failures and it will hurt, so plan ahead.
Also, regarding the new services that come online monthly: they will impact you. Today, your portal interface works one way; tomorrow, it is a different process. You spent weeks developing and fine-tuning a custom deployment script against their API – and it just broke because a new monthly release changed their API. They will try to limit the impact on you but growing quickly will lead to things being missed and impacting you unexpectedly.
They also know that once you join their team, it is hard to leave. If your Dev team wrote software using AWS Beanstalk, you reaped the benefits of ease of deployment and manageability. This is the promise of the Cloud being delivered. Keep in mind, though, that if you find issues or a cheaper price somewhere else, you are now locked into that automated Beanstalk solution and cannot easily unwind your integration with it. Plan accordingly.
Packing up is always hard to do
I moved into a new home a few years ago, and I vowed to avoid doing that again for a very long time. It was very painful to pack up everything, move it and unpack again. You will find the same pain when moving in bulk to a cloud solution. Transferring double-digit terabytes over the wire can take days or weeks (and that is if you have a dedicated line). What if your network connection is already saturated with normal daily usage? Worse yet – what if you must transfer a petabyte? Big providers may offer snail mail solutions to seed your data, but that will still take weeks or months to complete.
Things get harder when you consider migrating all your services. A Cloud architecture is not the same as your on-premise design. You will have to redesign some of your architecture to protect your services and data in a way you never did on-premise. If you don’t, remember that their sorrow will become your own. Planning the move is important.
Oops, I didn’t see that coming
There is nothing worse than buying a home, moving in, and then suddenly having something break. In our mad rush to the Cloud we may overlook important details. Redesigning a poorly-designed Cloud network will require rebuilding everything that sits on top of it.
Do you know which laws from which country, state, or municipality apply to you and your data? It is still a little bit like the Wild West out there, and lawyers are still busy working out the details about what applies to whom.
The real killer to quick migrations to the Cloud is the unexpected costs. IT admins have grown to be lazy with on-premise deployments. There is plenty of RAM and CPU available so right-sizing guests was never a priority. Your oversized guest now costs a fortune in the cloud. Even if you try to project costs ahead of time, you will find that online cost calculators are confusing. Moreover, you will likely miss calculating for resources that you will have to pay for. The options are endless in Cloud, and the monthly cost will vary based on your workload. Planning will be a challenge.
Winner, winner, chicken dinner
There is a theme though. The only way to achieve the potential that the cloud offers is to plan and plan some more. Accept the fact that some workloads and systems are just not ready to be moved to a Cloud provider. Realize that the flexibility of the Cloud does not mean it is infinitely flexible and that there are limits. The Cloud can be one toolset leveraged by your organization to achieve huge benefits, but planning cannot be an afterthought. Know your goals. Know your requirements. Estimate your costs and make sure to have buy-in with management. Be realistic about the use cases, benefits, and risks of Cloud. That way, you can swallow the tough pills you must and avoid the pitfalls of many.
The Fulcrum Difference
At Fulcrum Technology Solutions, we differentiate ourselves from other technology- and business-consulting firms with a unique guarantee: when you hire Fulcrum, we commit to finish the job. Whether working under a time-and-materials contract or a cost-plus arrangement, we will not leave until we’ve delivered exactly what we said we’d do. Our word defines us, and motivates us to give you the service that you deserve!