Levers for cost control on cloud

(These are totally my personal thoughts!!)

I have been reading about different viewpoints related to future of computing and cloud not able to live up to cost expectations.

- While one area that gets clearly articulated and talked about is cost of code quality. How optimally a code is written and/or whether the code is using right paradigms that enables efficiency are very important for digging deeper into cost and lever for cost effectiveness.

There are other levers that can make a considerable impact on cost:

- Differentiated strategy for data storage and data in motion (processing of data), organizations keep of storing data without identifying what all is necessary and what can be discarded, when it can be discarded & who(services) should own (store) data. While the cost of storing data is cheap, the cost is high for processing data and transmitting data. E.g. in a payload each time whether the same information is being sent for data that can be put in separate call which execute on lower frequency or only if the value is changed. While this is hard to identify and control, it can be a great starting point for organizations looking for cost optimization. Also, there are other risks involved in creating multiple copies of same data.

- Processes around the cloud operations, it is very easy to spin up and consume resources in cloud so organizations need to have a well thought through strategy for what, when and who can spin up cloud resources. Ability to identity different available resource for same kind of scenarios/workloads and suggesting right choice need deeper understanding of evolving technologies and frameworks. It is easy to control when and who, the need is to do it right, such that when some cloud resource is needed it didn't take days to just get the approvals. When we did set up a process for new cloud resources in one of my previous originations there was lots of noise and pain but in 3 months we did make significant cost savings. Once we had sufficient improvements, the processes were decentralized for smoother execution.

While the ability to optimize cloud cost goes beyond the above 3 points and is driven by long term strategies of organizations that dictate whether organizations invest in IaaS, PaaS or SaaS or look for newer paradigms such as #infrastructureascode , #containers etc. Irrespective of the strategy the ability to write quality code (not just optimized for function but for system), ability to reduce size of data & data in motion and defining & implementing processes around cloud resources can be great cost savers.

#data #computing #bigdata #strategy #future #infrastructure