As a Cloud and DBA Practice Director, I get this question regularly. It’s something almost all IT Directors and CIOs will ask me. All of them have been told repeatedly that they need to go to the cloud. But of course, since they are the ones who are ultimately responsible for what happens during such a move, they are naturally cautious about how much it will cost them. Hence the question: On-Premise or Cloud, which is more cost-effective?
On-Premise or Cloud, which is best for cost vs performance?
I would say concerns about costs and cost vs performance represent the bulk of the discussions I have with them. IT managers know exactly what a buck spent on a physical bare-metal server gets them. In contrast, subscribing to a Cloud service where a virtual machine is created on a bare-metal server that is not hosted in an on-premise data center is way less concrete.
Despite the fact that cloud servers are “invisible” for managers, it’s important to understand that the main responsibility of every major cloud provider is to offer “best of breed” products. If they want to keep you as a client and want you to move even more systems into their cloud infrastructure, they have to deliver the best possible performance at a cost that is, at minimum, competitive with your on-premise infrastructure. Of course, one expects performance to come at a cost. My experience tells me, however, that most administrators won’t have to pay more for cloud than what they already pay for their on-premise bare-metal server. In fact, in all of the cloud installations we’ve done, it has ended up costing them less!
The cost vs performance equation is an important one and cloud providers have taken notice. I can say that the people at Oracle Cloud Infrastructure (OCI) have done their due diligence and have come up with prices that are definitely interesting. Otherwise, how could they expect to have you as a client?
Move to the Cloud for Scalability
Just as it is important for OCI to be aggressive in regard to cost, customers also have to do their homework (Just starting out on your cloud journey? I suggest you check out my blog “Adopting the Cloud – Where to Start?“). Have you ever purchased a bare-metal server that was double the size of your needs? Perhaps you had successfully asked for more budget in your annual budget exercise and had to spend it no matter what. Or maybe you wanted to plan for a possible peak in the application’s activity. Either way, you ended up spending more money than needed. This concept is very important. As you likely know already, scalability is one of the major benefits of the cloud, where it’s incredibly easy to increase or decrease the CPU capacity or storage of a Virtual Machine (VM) or to a database server.
Based on this, why would you allocate twice the CPU you actually need? This over-allocation might make sense in the on-premise scenario, but definitely not in the cloud and not if you want to be cost-efficient.
Size your Requirement
Your homework, as a future cloud solution customer, is to properly size your requirement. Not based on your on-premise server size, but on the actual usage of the servers. If you have a 16 CPU server that gets only 30% usage, plan a VM with only 8 CPU. This way your move to the cloud will be cost-efficient. Don’t forget, you can simply add capacity when it’s required. This is what we call “scalability”.
If you properly size your future environment on OCI, you’ll likely find yourself winning both ways: cost and performance.
But, Can You Prove to Me That Moving to the Cloud Will Reduce My Operating Costs?
I often get this as a follow-up question. The best example I can provide of the cloud’s ability to reduce operating costs is when using Oracle’s Autonomous Database (ADB). This is something I went over from a DBA workload perspective in my blog “Is Autonomous Database Oracle’s big leap forward? Compare the good old days”. For the administrator, it’s basically the same answer. In ADB, there are no manual tasks for general database management, installation, tuning, or backups. Also, there is no need to apply patches, no need to recover the database, as Exadata and RAC proactively provide a very high level of protection. the result is a safer environment, reduced downtime, and countless man-hours saved.
Not everyone is ready to make the leap to Oracle’s flagship cloud product, but it doesn’t mean you can’t reduce costs. For those who decide to start smaller and simply use Oracle Cloud “Infrastructure as a Service”, here are a couple more examples of Operational cost reduction.
Using “Infrastructure as a Service”
First of all, not having to build infrastructure is cost saving in itself. Oracle Cloud Infrastructure delivers one, and you simply have to use it. You’ll cut on labor cost for the installation of the new racks, bare-metal servers, cabling, SAN, switches. Essentially, any component that makes up your infrastructure. Here are two scenarios this can apply to:
Sysadmin manages a Linux Server, load-balancer, SAN Storage, etc.
The cost reduction comes from not having to manage the infrastructure behind these Cloud Services. No need to change a broken disk, broken network card, or any other components that make up the infrastructure. It is entirely Oracle’s responsibility.
DBA manages Oracle database
The main cost reduction here (as mentioned in Autonomous blog) is that Oracle Database’s provisioning process is way faster than it used to be. Also, OCI provides features to help DBAs with their day-to-day tasks. In particular, with two critical activities every DBA must deal with: RMAN backup and the recovery and patching process. These are simplified, automated and entirely managed by Oracle and from the OCI Console.
Less Effort for your Team
These 2 examples illustrate how Oracle Cloud Infrastructure helps your operational team. Firstly, by managing the entire infrastructure and secondly, by automating different tasks as much as possible. Your team reduces efforts on these fronts so you can cut your operational cost.
Will I Lose Control of My IT Budget if I Go to The Cloud?
“Au contraire” you’ll have better control over your budget. Every year, you have to plan major investments in your yearly budget. Every year, you ask your team if they plan to purchase any equipment. This usually goes well, but say you are currently using 80% of your entire storage solution and unfortunately, because you had to invest somewhere else, you don’t have enough money in your budget this year to buy extra storage capacity. You will probably have to ask your team to clean up storage here and there and to do whatever they can to “find some space”. I’ve seen places where this happened, and management asked to reduce the number of backups. Not ideal.
By using a Cloud solution, you have access to infinite storage capacity. The beauty of it is that you’ll only pay only for the capacity used by your backups. No investments, no need to manage this capacity on-premise, simply use it and pay for it.
This was just an example of why you won’t lose control of your budget. Here’s another: If a major project goes on hold for any reason, simply shut down the project servers. You won’t get charged for them. Simple as that.
What is easier than simply paying for what you are using, compared to having to invest in on-premise equipment that you may not use at full capacity?