Many companies that migrate to Azure today, do a “lift and shift.” In other words, they take what they have on-prem and get the equivalent in the Cloud. But, is this too simplistic and will you be missing out on taking advantage of all of Azure’s capabilities?
The short answer, yes.
Scale out, not up when you migrate to Azure
When you’ve started thinking about doing an on-premises to Azure migration, you need to consider what you want from it. What can you do with your apps instead of just running another instance of a VM? Perhaps containerizing it will make sense. Or maybe a different delivery system will be better. Perhaps you could bring more functionality to it. This is scaling out, and lifting to modernize.
So the idea here is to tear down your processes and services and start applying new technologies to them — technologies such as serverless computing, containerization, or auto-scaling.
To understand what serverless computing can do in Azure, think about a customer who sends an email saying his bill is incorrect. You could build a serverless piece of code that would go off and say, “Okay, retrieve the bill, retrieve how we arrived at the amount, retrieve this, and send it to a customer support person.” You wouldn’t even be involved. All the dirty work will be done by the code, and when finished, that code shuts down.
Think of how many things in your business could be broken down into simple kind of actions like these and just get done. And then you could tie all those simple actions together. Think of the drudgery you could save.
Containers let you package your app with its libraries and other dependencies, just like virtual machines do. But they:
- Are much more lightweight than a VM
- Reside at host OS rather guest OS level
- Many can run on top of the operating system kernel
- Can start much faster and use far less memory than booting to an entire OS
- Are the fundamental building blocks for microservice architecture where all apps are developed as suites of independent services
Autoscaling lets you scale up or down as demand changes. For instance, perhaps your apps are in high demand during business hours, but virtually no one uses them at night. Perhaps you’ll need 10 instances during weekdays, but only a couple on the weekend. Autoscale to fit those needs, and in doing so, pay only for what you actually use.
Instead of just lifting and shifting during your on premise to Azure migration, consider what using it as a platform-as-a-service means. You can shove the management, shove the monitoring, shove the patching, shove all that ugly stuff that no one really wants to do that cause service interruptions and late nights and all that, and just let the platform manage itself. And the incremental cost is just nil.
Automate and modernize. Don’t just lift and shift.
The Azure migration guide: The process of on-premises to Azure migration
Don’t just jump into the cloud. Plan your Azure migration process.
1. Assess what you have
In order to figure out what your ROI might be in a move to the Cloud, look at what you have now and what you own — both infrastructure and software. What are your licensing costs? Power costs? Space utilization costs? Staffing costs? Categorize everything and determine where your biggest savings will come from. Perhaps the server space you now have is needed for new employees for your growing company. What will you save by not needing to construct a new space? Or perhaps you’ll be able to eliminate some IT staff who currently do all the patching, backing up, and server maintenance.
2. Where will you find benefits?
Think about what’s holding you back now. How can you get past that. Do you want benefits to come out of performance, cost, or reduced risk? Look at it in kind of a matrix, see what scores the highest in benefits and focus your energies on that.
3. Determine your ideal time
You know the scenario. Next year, you’ve got a lot of renewals coming up, or the SANs is going to wear out, or you need new servers. That’s a big CapEX coming up. This could be the ideal time for an on-premises to Azure migration. Instead of laying out $50,000 to $100,000 in the upcoming year, defer some of it. Take some of the low-hanging fruit — think your Exchange server and Office software — and start your migration there.
4. Take the T-shirt approach
T-shirts come in sizes small, medium, large, and extra large. So, a small modernization might just involve supplying a simple application or platform — in other words, a tool-based migration implementation. A medium-size would involve applying for an application or platform, which requires some customer interaction. An extra large modernization could be the migration of the PeopleSoft behemoth, a very complex, custom, and multi-tier application.
Once you’ve decided what you’re going to migrate, you begin preparing. You do some due diligence and know what needs to be done.
The next step is validation. You need to assess whether your on-premises workloads are ready for migration, and if not, make sure they are before migrating over — otherwise, you can make some critical mistakes that will be hard to undo, and you could end up falling all the way back to zero.
6. Azure server recovery
Azure Server Recovery came out as a disaster recovery tool, where your servers are mirrored into Azure, kept in sync, and can fail under certain conditions. But the tool has actually turned into the primary way you migrate your servers and data into Azure. It manages picking up that image and that data and everything and replicating it into Azure 100%. It’s not you going in and trying to move this. This makes life a bit easier.
The benefits of on-premises to Azure migration
You’ve heard all the hype about the cloud. But let’s look at some concrete benefits that could be yours when you doing an on-premises to Azure migration.
Let’s say you’ve got multiple locations, with some functions duplicated in each facility. Think how centralizing everything makes life easier. You can sit at the control panel and provision, move, do whatever you want across the infrastructure that could be anywhere in the world. And, you can do it from anywhere.
Reduced labor cost
If you’re like many companies, you’ve got one IT guy focusing on SANS and storage, and another on networking, while another focuses on yet another part. It’s very verticalized and you’re very dependent on each of those people. When you migrate to the cloud, much of those functions can get consolidated down and one person can play that role.
Pay for what you use
You can spin up an instance, test it, find a few issues, spin it down, and fix it — for minimal cost since you only pay for what you use. And the fixing part is on your time and your equipment. Then you can spin up again, test again, and spin down again. If you need more capability on, say a Black Friday, simply scale up, and when the day’s over, scale it back down. Microsoft says that you should be able to see 30% cost reduction.
Be quick or get left behind
One of Azure’s exceptional benefits is speed — speed of deployment, speed of operation, and speed of scalability.
Think outside the box
Ready to harness the full capabilities of the Cloud?
Great. But, if you’re concerned about downtime, or putting your data at risk, or losing some of it, talk to us, The Azure Cloud Experts. Basically, we’re going to seed everything into the Cloud well in advance. You’ll have two copies. You can look at it. It’s going to be complete. Give us a call today at 858-777-6980.