Conquering Cloud Costs: Part 3
- FinOps and Reserved Instances Made Easy
Post date: July 26, 2019
Cloud Operations Series
This is a 7-minute read and is part 3 of our 3-part blog series on managing your cloud costs.
In our previous blog posts for the “Conquering Cloud Costs” series, we discussed important topics concerning exactly what the biggest cost issues are and how to remedy them. In the first post, Understanding Cloud Costs from A to Z (link), we talked about understanding the basics of cloud costs and how public cloud services charge customers. In the second post, Reducing Instance Waste (link), we discussed the specifics of cloud expenditure and how reducing instances can make a huge impact in overall cost-cutting efforts. If you haven’t had an opportunity to read these two previous posts, you can check them out here:
Our final post for the series on conquering cloud costs introduces FinOps. We will focus on the most popular big cloud offerings created to save customers money including reserved instances (RIs) and committed use discounts (CUDs). We’ll take a look at these plans so you can decide whether you can save significant amounts of money for your organization by adopting one of them, or if switching to a newer cloud offering, such as Ankr’s Cloud Platform with flexible pricing plans, might be a better choice for your team.
Intro to FinOps in the Cloud
Evaluating your reserved instance or committed use discount options to optimize your cloud costs and performance is more important than ever. With cloud costs raising daily, it’s particularly vital to your cost reduction efforts to thoroughly evaluate your current plans and then make purchases or modifications that will give you the best bang for the buck.
Your company’s cloud is constantly growing and changing. Optimizing costs with reserved instances or committed use needs to be an ongoing practice with your internal finance, tech and business teams, your external tech partners, and your cloud vendors.
This kind of multi-team coordination is part of a growing practice known as FinOps.
AWS and Azure have both long offered reserved instance pricing. Google added their version of a cost savings plan in 2017 and named it Committed Use Discounts (CUD). This plan gives customers better pricing when they commit for longer than one or three-year periods. Payment, scope, tenancy, and other terms also drive pricing.
We will give a high-level overview of:
Deeper understanding of Reserved Instances (RIs) and Committed Use Discount (CUDs)
Contract Types: Standard vs. Convertible – Standard gives you greater discounts and you can sell them in the marketplace (of course Amazon has a spot market!), though you can only modify the RIs you buy. With Convertible you get more flexibility with the ability to exchange RIs for new ones (of similar value), though understandably you then cannot sell them on the spot market.
Payment terms: Reserved Instances provide three payment options: All Upfront, Partial Upfront, and No Upfront, and can save you up to 69% over On-Demand rates (when used in steady state).
Contract Durations: All Upfront and Partial Upfront Reserved Instances can be purchased for one or three year terms, while No Upfront Reserved Instances are only available for one year term.
Deeper understanding of Instances and pricing opportunities
Instance types: C, M, R, and X determine Memory, X, 2X, 4X determine vCPU. Most current generation are available for SAP / HANA.
Instance scope (Geos): Not all instances are available in all Geos. Need to specify Region or a specific Availability Zone within the Region, ie us-east or us-east-1b.
Instance OS: Amazon Linus (Opsworks Stack), Ubuntu 1X.04 LTS, CentOS 7, Red Hat Enterprise Linux (RHEL) 7
Instance tenancy: Instances can have Default (shared) or Dedicated hardware, and Reserved Instances can only apply to instances with the same Tenancy.
Optimize Cloud Computing Instance Costs
RIs and CUDs are only worth the expense and commitment if you are confident that it will be applied above the break-even point.This is where you have to match up your reserved instance purchase strategy with both your current and predicted Amazon EC2 usage.
Remember how reserved instances can only apply to the capacity of that instance? This idea is key to cost optimization with reserved instances.
If three instances run at once for eight hours, the reserved instance only applies to one instance and only eight hours, leaving 16 hours billed at on-demand rates, while running three instances one after the other, the reserved instances apply to all three for 24 hours with zero on-demand billing.
Imagine running eight different instances of the same type. For the purposes of this example, let’s assume the five hour pattern, which is repeated daily for the rest of the year.
At a glance, this charting method shows that there are always four instances running, with five running 90% of the time and six running 80% of the time. That means the following:
4 Reserved Instances (A-D) would have 100% utilization
2 Reserved Instances (F and F) would have 80% utilization
2 Reserved Instances (H and G) would have 60% utilization
Using this, you can select a waterline for your organization, or a savings level for your reserved instance purchases. If you selected an 80% waterline, then charting your use like this would instantly show you that you need to purchase six reserved instances to hit your savings goals.
How to Quickly Weigh Your Reserved Instance Options
Mature cloud infrastructures use thousands of EC2 instances, which means it’s time-prohibitive and cost-prohibitive to manually chart EC2 use against possible reserved instance use. With small systems, spreadsheets can be used to evaluate the information, but any growing cloud will quickly overwhelm the spreadsheet.
This is where cloud cost management tools such as Cloudhealth, Cloudflare, Cloudability, and others, come in.
Tying it all Together - Public Cloud Cost Management to Help Make Your Cloud Work For You
Ankr’s marketplace model supplies excess data center, mainframe, and device space to renters who need it on an on-demand, reserved, or spot basis. Everyone benefits from this model: Organizations and individuals who need compute are able to access it in more flexible formats. Organizations of all sizes with any significant excess capacity in small and medium data centers, server rooms, and mainframes located almost anywhere in the world are now able to recoup some costs. By renting out their excess compute capacity to Ankr and its customers, they can generate revenue that helps them offset sunk costs and boost equipment utilization rates. Surplus space is sold to customers who need services at rates often significantly lower than most big cloud offerings.
With their sharing model, Ankr’s monthly plans are lower than many big cloud plans. They are a notable significant 20 to 30 percent cheaper than up-and-comer and similar provider DigitalOcean, and a very significant 50 to 75 percent cheaper than AWS, Azure, and Google Cloud Platform’s current on-demand monthly plan rates. The available options and prices offer important savings and increased flexibility to customers.
Ankr Platform has many cost savings built into the design, including the ability to adjust to daily usage. Ankr’s technology is also faster, more secure, able to scale, and expandable; most importantly, it’s very easy to use and costs less. If immediate cost savings is your goal, Ankr is the cloud technology to adopt now.
In Part 1 of this series, we discuss why public cloud has become so popular, the cost drivers and introduce the tradeoffs between DIY on premise data centers vs cloud, the TCO , and how to best determine what you need from the different big cloud offerings such as on-demand, reserved instances, and spot instances.
In Part 2, we discussed how instance waste, typically driven by one or more of overprovisioning/underutilizing, misconfiguring AZs/GEOs, suboptimal contract and payment choices, can lead to overpaying big cloud providers.
There are many mechanisms in place with big cloud vendors that seem cost effective. Because costs are skyrocketing, AWS, Azure, and Google all must offer alternative price plans to their customers to stay competitive. It is time for other alternatives, but not just in price plans, new cloud platforms can offer more flexibility in pricing and service than the big guys. The costs aren’t a “potential” 75% but an actual cost savings over traditional plans. Now is a great time to try something new in cloud.