Purushottam Mohan Kshirsagar

Rakuten OneCloud services is a public cloud service with nine availability zones deployed globally, including in the US, EU, and Asia, to support and extend Rakuten’s eCommerce business in the respective regions.

DC Infra automation was achieved using Chef. Earlier, individual services were used to be deployed as stand-alone services. We improved automation by building a single automation code, reducing the timeline to construct a single availability zone. We have two different automation codes, one for the computing platform deployment and the other for network infrastructure deployment. The availability zone build starts with hardware procurement and rack and stack mount cabling as per the standardized architecture and continues to deploy computing Bare Metal (BMaaS) and VMware (VMaaS), DNS DHCP, IMS, Trident, monitoring, or basic services. We deploy the other services over the compute platforms like CaaS, LBaaS, STaaS, DBaaS, MONaaS, and EaaS OC portal.

Optimized the cost of delivery (COD) by left-shifting the work stack, automating Level 1 tasks, and streamlining operations by optimizing alert generation and people. This project aimed to reduce the COD or data center OPEX charges.

I reduced the COD by creating a service catalog to organize and categorize the tasks in Level 1, 1.5, 2, and 3 tasks in support. I also established a security and network operations (SNOC) team to left-shift all L1 tasks to this team by automating repetitive tasks and optimizing alerts to reduce unwanted tickets, which reduced the full-time equivalent count in each team (people optimization).

Using the effort-estimation method, a transition project plan, and a completed performance baseline, we reduced unwanted alerts by understanding the customers’ business and business criticalities. We achieved business needs and requirements by lowering the high-cost FTE count in run teams by recruiting low-cost L1 resources, causing a drastic reduction in operational costs.

I handled a SNOC team of 50 team members along with two team leads. Today, the work includes people management; shift escalation management; daily, weekly, and monthly performance reporting reviews; and two team leads.

Transitioned more than 200 servers with 30 to 40 interdependent applications. As the Move Group project manager, I transitioned the customer environment to the Wipro Data Center by preparing executive summary reports, including a project charter and project plans in layer one (L1), layer two (L2), and layer three (L3) phases of the project. The L2 project plans included minute details of the interdependencies of the application. By understanding the sizing, interdependencies, server types, and network communication with other servers, we gained insight into areas such as load balancers’ requirements and gathering the ports that need to be allowed to enable communication over firewalls. The L3 execution project plan was created using the L2 details by gathering the minute details and time required to complete each task. After transitioning with proper sign-off from the test, project managers hand over the environment to the datacenter run teams for 24/7 support.