Why zOpt: Part 4 of the series
What is "Cloud Debt" ?
All of us have known about "Tech Debt", a common phrase used for decades in the tech world.
Tech Debt is accumulated cost of choosing suboptimal hardware or software solutions, and the additional cost incurred for maintaining this suboptimal choices. This often results in increased difficulty in scaling the system for new business use cases and/or handling increased demand.
This Tech Debt is eventually paid off by refreshing the hardware or rewriting the software. In most cases, this means a significant effort, cost of getting your house in order. Tech Debt pay-off is directly proportional to the time one waits, longer you wait, results in additional accumulated costs and increased efforts.
In the context of cloud world, one is using hardware and software in pay-as-you-go model, in theory, it should easy to migrate to newer versions without any time lost. The accelerated pace of innovation in cloud world require cloud debt retirement much more frequently.
Tech-debt accumulation is slower because newer version of hardware and/or software is typically launched every couple of years.
Cloud debt is accumulated much faster because newer version of hardware and software is typically launched every year if not more frequently.
Hence, within 6-12 months you are probably running a generation older hardware/software. In addition, newer class of services are launched multiple times in a year.
Lets assume, you started using AWS S3 Standard storage a year ago supporting 100s of users and a few thousand transactions per day. In the past one year, your business has grown to 10K users, 100K transactions per day. In addition, AWS launched newer S3 storage classes that allows your workload run in better price performant manner.
AWS rolls out 3000+ features/services every year and a typical AWS customer use 30-50 AWS services. Within a few months, you could be behind a generation of hardware/software and missing out using a better suited service class for your demand seasonality.
For example, You started using RDS Aurora standard 6 months ago, whereas for your database transaction patterns RDS Aurora IO Optimized could provide you lower cost. Or, You started using M6 generation of EC2 an year ago, and now you could get price performance using M7 and M8 generation of EC2.
The rapid pace of innovation across cloud services accumulated a significant cloud debt in short timespan. This means you are using non-optimal cloud services for your business and paying higher cloud bills month over month. By the way, this phenomenon is equally applicable to Google Cloud & Azure as they are also innovating at break neck pace.
Choosing latest cloud service inline with your application architecture and demand seasonality could provide better performance, better customer experience while reducing your cloud spend.
How to retire Cloud debt
Retiring cloud debt is a continuous process as long as our demand patterns are evolving and cloud providers are innovating.
Step 1: Assess your cloud usage
Deploy and leverage cloud provider native tools such as AWS Cost Explorer or third party tools such as zOpt.ai.
Step 2: Identify cloud debt in your cloud environment
Assess the hardware generations you are using, map newer generations for your workloads and understand the price performance benefits
Assess the non native cloud services you are using, map cloud native services for your workloads and understand the price performance benefits
Assess the cloud native services you are using, map your workload usage patterns to other classes of cloud native services and understand the price performance benefits
Step 3: Prioritize cloud debt removal
List the cloud debt accumulated and options of retiring these cloud debts
Put together the efforts required for cloud debt retirement & associated cost
Prioritize the tasks based on cost & benefits analysis
Step 4: Execute the changes in your live production environment
Be sure that the intended change does not have side effects that could adversely affect your production environment.
Plan the task execution across stakeholders and multiple cloud accounts in tandem.
Step 5: Realize the benefits for latest cloud offerings
Step 6: Repeat steps 1 - 5 on periodic basis.
Simplify cloud debt retirement using zOpt.ai
Step 1: Assess your cloud usage
Available out of box with zOpt.ai
Step 2: Identify cloud debt in your cloud environment
Available out of box with zOpt.ai
Step 3: Prioritize cloud debt removal
Available out of box with zOpt.ai
Step 4: Execute the changes in your live production environment
Leverage human vetted automation and execute the changes across your cloud accounts with a few simple clicks. More importantly, All the zOpt.ai recommendations are proven with hundreds of AWS customers.
Step 5: Realize the benefits for latest cloud offerings
See the performance and cost impact on zOpt.ai's insights dashboards.
Congratulations, now you can retire cloud debt with utmost ease, which otherwise would have been multiple weeks of efforts for each iteration.