We recently had an opportunity to speak with the CISO of NuWare, Ajay Jamble. NuWare is a 25-year-old digital technology and infrastructure consultancy based out of New Jersey. Their clientele span some of the largest enterprises in Retail, Financial Services, and Healthcare amongst others.
Ajay is the Program Manager responsible for deploying Lucidity's managed EBS solution on NuWare managed databases. You can choose to either listen to the interview with Ajay by clicking on the podcast link or get a summarised version of the conversation here in the post.
Here's what we got from the sit-down with Ajay:
A: DevOps is still a manually managed process unless teams use a lot of automation tools and techniques. But as of now, the focus is to move from DevOps to NoOps, but that can only happen when we start to team up with other tools and services providers.
More than 45% of the infrastructure is manually managed.
A: Data is storage and data keeps expanding, so storage does tend to take a significant portion of cloud bills as the organization grows and expands.
30-50% of cloud costs end up being storage costs
A: We use EBS to store everything from log files to mounting databases and launching Kubernetes clusters. Smaller organizations prefer having smaller EBS volumes, but as the organization grows the EBS size also increases.
EBS makes up about 10-30% of total cloud costs
A: Predicting required EBS volumes is difficult unless you're an organization that has reached a level of maturity. But as a standard practice, we still deploy an N+N model, where if we need 1TB of data we tend to get 2TB to avoid running out of storage and having to raise change requests. This in turn wastes almost 50-60% of EBS storage costs.
60% of EBS costs are wasted as a result of the following reasons:
A: Unless you've migrated your databases to GP3 you're going to face performance issues with EBS.
Databases face frequent performance challenges
A: When there is a need for high performance, the choices within AWS are limited. This leaves teams with no choice except to launch multiple high-performance storages to get the computing power that is needed to run applications.
High IOPS disks are fairly expensive, this poses a lot of challenges for development and infra teams to get the computational power they need while keeping budgets in check.
A: Helping teams move away from the old N+N model of allocating storage would benefit infra teams plan storage better. An ideal solution would be a fool-proof way to ensure storage expands and shrinks without any manual interventions while keeping the applications on the storage working seamlessly.
Dynamic expansion is an ideal solution
A: Watching Lucidity shrink and expand storage without us noticing any changes to the performance of the application was one of the most exciting things for us to see. Another major benefit was getting a performance booster of 10x that negated our requirement to go for an IO2 disk and save costs while getting the necessary IOPS boost for our applications.
Postgres databases and other test data similar to the overall broad applications we generally use. We also ran stress tests and threshold tests to check the expansion and shrinking features. The DBA team ran simultaneously ran tests to ensure that the applications weren't getting affected in any manner.
Build a similar suite for FSx for Windows.
NoOps is the next logical step to DevOps.
Want to know more about how Lucidity can help save your DevOps team's time and effort? Click the 'Get Early Access' button in the navbar, and we'll get back to you at the earliest!