Jeyappragash JJ

    Jeyappragash JJ / Software Engineer, Ex-Twitter

    Jeyappragash helped bootstrap Cloud Infrastructure Management for Twitter. He built the team, and defined the product and technology direction for the team. The platform was designed as a multi-cloud platform that helps developers manage their services and applications. For operators, it provides detailed visibility to the infrastructure and how the tenants use their infrastructures. Previously he was a Distinguished Engineer at Motorola (then Google Company), leading efforts to build their Notification Infrastructure, their Software Upgrade services and a Prospective Search based Content Delivery Service, consumed by every single Android device that Motorola shipped.

    Jeyappragash graduated from IIT Madras with a Masters in EE. Jeyappragash holds 5 patents in cloud infrastructure and distributed systems space.

    Talk: IAM at Scale and on Hybrid

    In this talk we propose PADME, a set of standards to specify policies for access control of resources and a runtime architecture that can be used for enforcing access control at a fine-grained level.

    Complexity of applications architectures is increasing. Multiple architectures, from fine-grained services to monolithic systems, are required to coexists and environments range from in-house IT & private datacenters to public clouds.

    As monolithic services get broken down to smaller distributed fine-grained services, they are required to interact with the legacy systems. How does one define and control access to these dynamic resources in such heterogeneous environments?

    Resources protection is needed at different levels of the stack, from basic network to application data. As resources get more dynamic, the access pattern between multiple resources and users grows exponentially. With that complexity, managing access control gets challenging. There is a combinatorial explosion of resource access pairs that need protection. A brute force approach is expensive both in configuration complexity and runtime cost.

    Thursday 10th August 2017