Pradeep Kathail

Technical debt: past, present, and future (panel)

The term “Technical Debt” was coined over 20 years ago by Ward Cunningham in a 1992 OOPSL A experience report to de scribe the tr ade-offs between delivering the most appropriate — albeit likely immature — product, in the shortest time possible. Since then the repercussions of going into “technical debt” have become more visible, yet not necessarily more broadly understood. This panel will bring together practitioners to discuss and debate strategies for debt relief.

Collaboration and communication: growing and sustaining ultra large scale (ULS) systems

Mission- and life-critical Ultra-Large-Scale (ULS) systems are increasingly prevalent and networked in many domains, including business, aviation, communication, defense, finance, health, and public utilities. Such systems are often too complex for generally centralized methods to work well for such tasks as requirements discovery, development, system integration, test, deployment, configuration, operation protection, and evolution. Yet today we lack sound methods and technologies for distributing these tasks across large ecosystems of system production. What technical, legal, contractual, and cultural frameworks are needed to enable global partners with independent, sometimes conflicting agendas, to function effectively in the execution of such tasks? Can a competitive and collaborative distributed design ecosystem deliver value and robustness over time consistent with demands for quality, intellectual property protection, and other such requirements? Join this panel of industry experts and academic researchers who will share and debate their perspectives and lay out a vision for the future.

iRP: intelligent rendezvous point for multicast control plane

Multicast is widely deployed in data centers for point-to-multi-point communications. It has an established set of control protocols such as IGMP and PIM that has limitations around the lack of bandwidth awareness when establishing the multicast trees. This could leads to over-subscription of network links and packet loss impacting user quality of experience. Other existing multicast issues are around security, data flow efficiency and load balancing. In this paper, we design and implement a novel Intelligent Rendezvous Point algorithm, utilizing a controller-based architecture, to optimize multicast tree formation and to increase bandwidth availability in the fabric. The system is implemented using off-the-shelve commercial available switches. the proposed iRP algorithm maintains the creation, expansion and removal of source trees based on flow bandwidth and security requirements. The implemented system utilizes switches and data center fabrics that run traditional routing protocols and it does not require end points to change the multicast signaling as they can continue using IGMP. We implement the algorithm using controller-based system architecture and test the algorithm using Cisco Nexus commercially available switches. Testing results confirms that iRP Algorithm ensures that multicast flows are admitted to the fabric based on available bandwidth to ensure no over subscription on some of the fabric links. The system offers up to 50% increase in fabric capacity to handle multicast flows passing through the fabric. Finally, we confirm system ability to admit or deny sending or receiving flows based on security requirements.