Years ago, Filene’s department store in Downtown Boston had a “bargain basement” where clothing and items that weren’t selling in the main store got deeply discounted to move them out. It worked beautifully as people loved digging through bins for bargains. Believe it or not, the big cloud companies have a similar bargain basement, and you can learn to take advantage of it for your clients.
In the cloud case, it sometimes involves excess capacity that isn’t being used at any given moment in time. The companies would rather get some money rather than no money, so they cut the pricing dramatically. There is a catch though. When the company needs that capacity back, it will take it over and end the workload you or your client is running.
The upside is that AWS will discount your rate up to 90 percent if you are willing to assume some risk. There is also a way to mitigate that risk by using a tool like Spotinst, a startup that uses AI and machine learning to watch the cloud platforms and move your workloads to another bargain basement before you get shut down.
Azure has a similar tool called Low-Priority VMs, which launched earlier this year. Google’s version is called Preemptible VM Instances, which can save you up to 80 percent.
On the flip side, the cloud companies surely would prefer to have as little excess capacity as possible. One way to ensure that is to sign up customers for a specific period of time. MSPs may want to take advantage of steep discounts offered by cloud platforms to customers who make 1 to 3 year commitments.
MSPs may want to take advantage of steep discounts offered by cloud platforms to customers who make 1 to 3 year commitments
AWS will give you up to 75 percent off of retail pricing for signing up for their reservation product called Reserved Instances. Azure has a similar offering called Azure Reservations, which offers a discount of up to 72 percent off of full retail pricing. Google’s version is called Committed Used Instances, which provides a 57 percent discount for most machine types or custom machine types and up to 70 percent for memory-optimized machine types.
Cloud pricing tends to be a moving target that’s hard to nail down, but if you shop around and learn about the different options, you can present the different tiers to your clients. If it’s a low priority workload, maybe the spot instances will suffice. If it’s higher priority and they want cost certainty, maybe reserved instances will be better.
Regardless, you need to understand the pricing structures and what you and your clients are paying because you certainly don’t want to pay more than you have to. If you can use discounted services, you should consider it.
Photo: William Potter / Shutterstock.