Before we start, here is a quick refresher on the 3 types of cloud solutions, just in case you need it!
Someone asked me the other day why any company would move their IT infrastructure to the cloud. They could see the benefit of SaaS platforms, like Office 365 and ServiceNow since those solutions remove the OS and software management layer nearly completely and are just extremely easy to setup and use. They couldn’t wrap their head around moving other workloads to the cloud, especially IaaS. For example, why would you move some proprietary internal software or Windows file shares to a cloud solution when you’ve built up your datacenter kingdom? Over the years, companies have invested a lot of money in on-premises infrastructure, why change? Let’s dive right in there!
Heavy Capital Investments
Depending on the size of your company and the workloads that you run, your IT infrastructure could be massive. Networking, server and storage companies have done a fantastic job of optimizing workloads on-premises, but at a cost.
Let’s look at a robust virtualization environment.
Your company has experienced sustained growth for several years and has had an initiative to get over 80% virtualized. It is now several years later and you’ve reduced the amount of physical servers, thanks to blades, tiered storage, and advanced networking across datacenters that allows you to do some amazing things with DR. The environment is huge, but runs well with the latest automation tools.
Phew. That’s a lot of capital you have invested in. Accounting is less than thrilled because they are still depreciating all those assets, but you have saved money in consolidating physical servers. Your CIO has no idea that your current utilization is only 70% because you over-purchased, but at least you have capacity. Plus, there are 4 environments for every critical production workload – dev, QA, support, sandbox. Most of these servers sit online and idle for 80% of the time. If they were offline, your utilization numbers would be closer to 40%.
Every year, you look at these numbers, talk to business application owners and try to predict new capacity, but new unexpected projects always come up. You always try to plan 20% more just in case to handle the additional phantom workloads.
A Better Way
Here’s where the cloud comes in. Instead of all that infrastructure, capacity issues and depreciation, you could utilize a cloud solution for some of those workloads. While capital expenses are difficult to predict and more difficult to assign to business units across a shared infrastructure, operating expenses are extremely easy. With a cloud service, you have the flexibility to only pay for what you use and scale up or down depending on your needs.
Here’s a great example.
Let’s assume you work for a national company and your resources are accessed only 12 hours a day, 5 days a week. If you had on-premises infrastructure, it is likely online for 24 hours a day in a data center with constant power and cooling. That is 720 hours of work per month (not including all that other stuff). If you really only needed 12 hours of capacity per workday though, that works out to be around 264 hours, That’s 50% savings, plus potentially hundred of thousands of dollars in savings from hardware costs and depreciation. Lastly, all those non-production servers can also be moved up and only turned on when they are needed, saving you even more money!
The cloud gives you this easy scalability.
When is a Good Time to Start Your Journey?
There’s two ways to look at this. If you are considering a massive all in cloud solution, which to me sounds awfully scary, a good time to look is during your network, storage and server refresh cycles. I would imagine that this is very hard to plan around though. Is your company regimented enough that you swap your entire datacenter out every 3-5 years and gain immediate savings? Probably not.
There are other ways to start on your cloud journey. SaaS based applications are probably easiest and most often the entry point without IT possibly even realizing they are in play. For example, do you use ADP for your paychecks? That’s really SaaS payroll right there!
When we are talking PaaS or IaaS though, all those non-development workloads are a great place to start. Think about what you do when provisioning a development VM for someone. Typically, you give them a lower spec machine that just sits there on all the time. You can get these workloads easily moved into a cloud service, like Azure, allocate the full specs to match your production environment and have your devs hit it. There are two benefits here. First, you are freeing up on-premises capacity for production workloads and second, you can now spin up and down the VMs and only pay for the usage (down to the minute in Azure)!
Scalability is Key
When cloud solutions first came to light, there were a number of cost saving measures offered. Among those were lower staffing considerations, which is always a tough topic. Staffing savings might be there, but it is more likely that your existing IT staff will still exist, but need a different skillset.
There are other features that the cloud can offer for your workloads. The cloud can easily provide backups and recovery. Another way is how cloud providers have worked with ISPs to provide direct connections into their services. Now, this does come as a cost, but it does have a lot of advantages for performance.
Security is another consideration. All I will say is this. Consider what you can do internally and then compare it to a company whose business interest is in protecting your data because their revenue stream depends on it. Microsoft has even gone so far as to fight the US government over it.
Putting all this aside though, PaaS and IaaS cloud really helps IT organization plan and chargeback to business units much easier than it’s ever been done before. This centers around scalability, but also capability. Imagine being able to move and adjust quickly as your business demands it. That’s what the cloud offers.
What’s the Catch?
As easy as all this sounds, a sound automation and operation model should surround the non-SaaS cloud in order to take advantage of all the cost savings. It does require a strong partner likely to get you there. This is key to a successful cloud implementation, but should be a part of any successful on-premises implementation as well.
This is why Microsoft’s hybrid cloud solution is so compelling. It focuses on getting the pieces to automate and manage all workloads first and then leveraging the cloud for what you can using those same tools and principals. There’s no need to rip out expensive infrastructure and go all into the cloud, but it can help you save money as you refresh environments.
The cloud can be a gradual journey that allows you to streamline existing operations before you move any workloads. Develop processes around what IT does well – enable business success through technology and then determine if they cloud is right for your workloads.