Codes for Unexpected and Expected Loads in Distributed Computing Systems
Explore a technical lecture by Distinguished Professor Emina Soljanin from Rutgers University examining how erasure coding can optimize distributed computing systems to handle unpredictable data access patterns. Learn about the concept of access service rate region and its crucial role in designing stable distributed systems that can manage varying mean loads. Discover how mathematical tools like combinatorial optimization on graphs and finite geometry address challenges in coding theory for distributed systems. Gain insights into maximizing concurrent data access requests with fixed resources through data object replication based on popularity and access volume, particularly relevant for edge computing applications where user numbers and data interests fluctuate significantly. Delivered by an IEEE Fellow and former Bell Labs researcher with extensive expertise spanning distributed computing and quantum information science, understand the intersection of coding theory and practical system design for modern distributed computing challenges. Read more