Faglig kontaktperson er professor Otto Anshus.
Kart over Campus: ( https://uit.no/campustromso )
Alle interesserte er hjertelig velkomne.
Vel møtt!
----------------------------------------------------------------------
Date: Friday April 10th 2015, 12:15
Place: Store Aud, Realfagsbygget, UiT
Title: " Pub/Sub: Challenges of Google/Cloud Scale "
Lecturer: Dr. / Staff Software Engineer Haakon Ringberg, Google
Abstract
The increasing number of workloads that move to cloud computing environments ( e.g. , within containers ) need a reliable messaging system for intrasystem and intersystem communication. The most prevalent such messaging paradigm is possibly PublishSubscribe (pub/sub), and pub/sub systems have flourished in popularity and are now an essential messaging product for cloud providers to offer. Google has Cloud Pub/Sub , Microsoft has the Azure Service Bus , Amazon has bothSQS and SNS , and there are numerous other open source and enterprise offerings [1 , 2 , 3 , 4 ]. Pub/Sub systems are popular because they offer a decoupling between the senders and receivers of data, provide a general API that is applicable across a range of use cases, can be architected to scale horizontally, and may offer other appealing features ( e.g. , discovery, backlog management, flowcontrol, retry). The first half of this talk is an introduction to the pub/sub messaging paradigm.
Pub/Sub will also be used as an example system in order to discuss challenges in providing a system as a service for a cloud computing provider. Specifically, while pub/sub is a simple systems idea, it is challenging to design/manage a pub/sub system that can run as a service, which supports arbitrarily shaped workloads. The second half of this talk is an introduction to these challenges.
Short bio
Haakon Ringberg is a Staff Software Engineer at Google where he is the Tech Lead for the pub/sub team and product. The product has recently been released as Google Cloud Pub/Sub. Haakon received his Ph.D. in computer science from Princeton University working with Professor Jennifer Rexford. His doctoral dissertation pertained to efficient methods for detecting unwanted traffic on computer networks and introduced a novel algorithm and system for enterprises to collaborate to detect unwanted traffic in a privacy-preserving manner. Haakon has published papers in SIGMETRICS and INFOCOM among other conferences. He lives in NYC but grew up in Tromsø.