There is still quite a lot of chatter about cloud these days and it doesn’t seem to be showing any signs of slowing down or getting less complex
“What’s your multi-cloud strategy?”
“How about hybrid cloud? What’s your plan?”
“How about your on-premise hybrid multi-layer hyper-converged cloud?”
The variations are dizzying. So we’ve decided to keep our move to cloud technology a little simpler to start. While we’re finding our way slowly but steadily with targeted projects, we are learning what works well and how best to get there through these experiences.
This approach takes planning at a strategic level in determining cloud’s place in the overall technology portfolio as well as orchestrating among multiple technical teams in pursuing targeted initiatives. Significant time is also needed to implement and assess results and progress made.
We began with moving email to the cloud – not an earth-shattering endeavor but it’s a start, and email maintains a significant production workload. There is a mounting track record for email migrations to cloud and at this point it’s most likely considered low hanging fruit for most organizations.
Currently we’re also in the midst of moving our analytics platforms to cloud. That’s a little more complex than email. So we’re migrating in phases and iterating as we go along. A few positives we’ve experienced so far in migrating to cloud:
- On-site consulting assistance from our cloud platform vendor and a local consultant helped us avoid technical “potholes” and accelerated our staff’s learnings;
- Things that might have taken us months to deploy with legacy on-premise technology took weeks with cloud;
- Cloud is scalable and it works. We increased some data workloads 300x with no additional effort;
- The burden of supporting the underlying technology is removed from our analytics team, allowing focus on advanced data modeling and analytics functionality;
- There is a strong user base and support community from both our platform vendor and other health systems.
That said, we are still determining the long-term cost of moving all of our analytics workloads to cloud. We just don’t know enough yet to have that answer. I hear a fair amount of negative comments about the cost of cloud and how it’s more expensive than on-premise.
But I haven’t really seen any long term side-by-side comparisons to validate those remarks and wonder if that’s really true or just the experience of some that haven’t properly managed their cloud infrastructure and usage. Heck, your cable bill is a great example of how an unmanaged service can get out of control.
We’re modeling our experiences each time we move a workload and use that knowledge to develop a profile that helps us plan for the next opportunity. We are also now facing the tough work of developing migration plans for all of our existing analytics infrastructure. This is not a trivial task for sure and we’ll learn along the way.
My advice for anyone that seems to be in a “fog” about the “cloud” is to keep in mind that it’s really not a new whiz-bang just-invented technology. At its core, it is an updated version of traditional economy of scale CPU and storage sharing that’s been around for decades, with a variety of new bells and whistles for sure.
Proper planning, oversight and management go a long way – just like these elements have always been the key underpinnings of successful technology pilot, rollout and maintenance.
My ultimate recommendation is to use these cornerstone approaches for cloud migration just the way you would any other technology project. And lastly, don’t bite off more than you can chew. This will help you avoid the pitfalls that have always been in play with Information Technology.
Share this post if you enjoyed! 🙂