Wu & Chen Auditorium (101 Levine Hall)
CIS ColloquiumJon Moore
"Not-So-Big Data After All: Managing Reference Data with Sirius"
Many interesting “reference datasets” now fit in a single commodity server’s RAM, with more on the way as main memory sizes continue to grow. The open source Sirius library allows developers access to this data in native datastructures while managing the distributed systems heavy lifting of replication and persistence.
In this talk, I’ll describe what “reference data” means so that you can recognize this common use case in your own applications, highlighting problems that Sirius is a good fit for (and those that it isn’t). I’ll describe the overall architecture of a Sirius-based application—-the high-level concept of how it works.
Finally, I’ll use a case study from Comcast as an example where Sirius is powering applications serving tens of millions of customers.
This talk covers the highlights of a paper we published in the USENIX Annual Technical Conference this year.
Jon Moore describes himself as being equally comfortable giving architecture talks and personally writing production-ready code. A Technical Fellow at Comcast Corporation, he runs the Core Application Platforms group that focuses on building scalable, performant, robust software components for the company’s varied software product development groups. His current interests include distributed systems, hypermedia APIs, and fault tolerance. Jon received his Ph.D. in Computer and Information Science from the University of Pennsylvania and currently resides in West Philadelphia.