How to optimise cloud operations
Wed 28 Oct 2020 | Jeff Collins
With a few smart moves, every company can become a cloud operations optimiser.
There’s a fine line between cloud spend and cloud sprawl. Most companies today are using cloud technologies to power their most important products and services, communications and collaboration, but it’s easy to cross the line from spending smartly on cloud services to spending unnecessarily
Gaining control of cloud operations and cloud spend should be a goal for every organisation, big or small. The good news is, it’s not as tough as you think. With a few smart moves, every company can become a cloud operations optimiser. Here are a few tips:
Tip #1: Gain Complete Visibility Into Your Data
It’s all about gaining a comprehensive view of all your data, pulling correct metrics, ensuring tagging best practices are followed, and analysing/interpreting what is going on. Many of the cloud optimisation tools out there don’t pull the necessary information (actual RAM usage, for example) to accurately make meaningful recommendations.
Compute resource data is only part of it; having a deep financial understanding around when to make RI/Savings Plans purchases and EDP/EA negotiation are just as important. You don’t want to pay for too much coverage or too little, and ideally you don’t want to make those decisions before you know your environment has been optimised.
Tip #2: Know your Applications
Cloud providers offer multiple financial methods of procuring resources such as Savings Plans, Reserved Instances, Spot resources, EDP/EA Plans, etc. Knowing how your users are interacting with your applications is key to understanding which financial option will make the most sense and achieving maximum savings.
For example, a client has an application that is very spikey in utilisation and is only accessed at certain hours during the work week. Purchasing a Compute or EC2 Savings plan or RI may not be as cost effective as applying auto-parking policies or utilising Spot resources (or a combination of both), since the application doesn’t need to be on all of the time.
Tip #3: Non-Traditional Cloud Services
Frequently, clients migrating to the cloud will simply perform a 1-for-1 migration and carry with it the traditional three-tier architecture (Web/App/Database). Getting out of that mentality can be a struggle for many clients, but by having awareness of what is going on in the cloud environment at the application layer can help formulate a strategy for that over time.
It can also help transition that mentality away from traditional architecture into more cloud-native services (containers/serverless, for example), which can provide cost savings, improved performance and open the door to more advanced services that can be taken advantage of in the future, such as machine learning and AI.
Mistakes to Avoid
Too often, companies believe that their cloud environment is already optimised. A typical cloud provider rolls out approximately 20-30 new features and updates per week. Multiply that by three for the top cloud platform providers (AWS, Azure and Google Cloud), and you have a daunting challenge of keeping up with everything. This takes the client away from doing her/his most important job – running their business.
Companies also have a tendency to not think about optimisation or application modernisation when deciding to move to the cloud. A lot of clients will simply look at their compute requirements that are being utilised in their data center and build the exact same thing in the public cloud. While this will get them to the cloud quicker in most cases, there tends to be a lot of waste when performing this method.
The sooner you can start thinking about optimising the better. And, finally, too many companies take a traditional data centre operations mentality to the cloud. The cloud continues to offer more and more advanced services that can greatly benefit a business.
Taking a holistic approach to optimisation and application modernisation leaves no stone unturned, typically generates enough savings to pay for itself, and, once implemented on a regular basis, will ensure that application performance meets demand. This is especially effective for clients with large cloud spend and dynamic environments.
The best way to optimise cloud operations and spend is by engaging a Managed Services Provider (MSP) and realising that there is no single tool that can provide the type of analysis most companies need. Optimisation requires both in-depth technical and cloud financial knowledge.
It also requires collaboration between the MSP and all of a company’s business units to gain an understanding of their needs and application usage. Once all the data has been collected and interviews performed, recommendations can be delivered. It’s then up to the client, in collaboration with the MSP, to implement those recommendations to achieve true optimisation and ultimately realise cost savings.