If you’ve been in technology for a while, you’ve probably had to go through a hardware refresh cycle at some point. These cycles usually meant taking existing hardware, doing some capacity planning exercises and setting out to buy new hardware that is supported by the vendors. This process was usually lengthy and made CIOs break into a cold sweat just thinking about paying for more hardware, that’s probably just meant to keep the lights on. Whenever I first learned of a hardware refresh cycle, my first thoughts were “Boy, this sounds expensive!”
I’m sure that hardware vendors loved to hear about hardware refresh cycles. Sales teams knew they would probably make a new sale and better yet, the customer probably called them up to do it. Very little work on their side,but things are much different now due to so many customers moving to the cloud where the hardware is not the customer’s concern any longer.
Cloud Refresh Cycles
So hardware refresh cycles are dead now, right? I mean, you’re in the cloud now, and no longer need to worry about this kind of nonsense. Well, sorry to burst your bubble here, but they aren’t quite dead just because you’re in the cloud. Now, while capacity planning becomes much easier, and calculating budgets for new hardware is much simpler, you still need to review your infrastructure and make sure you’re up to date, but now for an entirely different reason. Saving Money!!!
Oh! I have your attention now huh? I thought so. In the cloud world, upgrades can save you money. Take a look at a few examples here just from the AWS platform.
I used the simple monthly calculator provided by AWS and looked up the price for a Linux instance in US-East-1 for m3.large, m4.large, and m5.large. These are similar sized instances but three different generations of hardware. M3 instances being the oldest of the three and M5 the newest of the three. What you’ll find is that the size of the instances is virtually the same, but the hardware that comprises the instances is faster. That makes pretty good sense, but also notice that the prices of those instances gets cheaper as it gets newer.
|Instance Family and Generation||Instance Size||CPU and Memory||CPU Speed||Pricing|
|M3||Large||2 vCPU / 7.5 GB Memory||High frequency Intel Xeon E5-2670 v2 (Ivy Bridge) processors*||97.36|
|M4||Large||2 vCPU / 8 GB Memory||2.3 GHz Intel Xeon® E5-2686 v4 (Broadwell) processors or 2.4 GHz Intel Xeon® E5-2676 v3 (Haswell) processors||73.20|
|M5||Large||2 vCPU / 8 GB Memory||2.5 GHz Intel Xeon® Platinum 8175 processors with new Intel Advanced Vector Extension (AXV-512) instruction set||70.28|
So let me reiterate that point one more time. If you upgrade from older, slower hardware to the newer faster hardware, you’ll not only be gaining performance, but you’ll be saving money while you do it. Pretty neat huh?
Let’s look at one more example. Here we show the price of S3 standard storage vs S3 Reduced Redundancy storage. Now this isn’t a generation thing, but Amazon isn’t putting a lot of effort into the Reduced Redundancy Storage option and would prefer to spend time on the standard S3 storage that is used for most workloads. Notice here that as time as gone on, Reduced Redundancy Storage is actually more expensive than the standard storage even though it provides 7 Nines less durability than standard.
|S3 Storage Type||Durability||Size (GB)||Puts||Gets||Price|
Now that you know this, it is a good idea to have an action plan and pay attention to new releases that come out in the cloud. Sure you can keep running your workloads without messing with them until the cloud provider discontinues your generation of hardware. But you could take advantage of these new performance enhancements from better hardware while saving money along the way. Set a schedule to review your infrastructure and determine if and when you need to upgrade your own cloud infrastructure. Maybe this is a good time to leverage a partner to help education you on the changes that are constantly happening in the cloud.
How much money are you wasting?