skip to primary navigation skip to content

Cambridge Systems at Scale (CamSaS)

Department A-Z



The Musketeer workflow manager runs workflows across multiple back-end frameworks, choosing the most optimal one automatically...

Read more


With QJump, we are bringing guaranteed latency communication to data centre networks...

Read more


Legacy OS abstractions from the 1970s do not scale to modern "warehouse-scale" data centres. DIOS is a new operating system designed for scalability and efficiency...

Read more


Using CamIO allows applications writers to focus on the important aspects of writing application logic whilst ignoring the complexities of I/O management...

Read more


Modern data centres, web applications and big data analytics require systems software to operate at unprecedented scale.

CamSaS is our initiative to build a systems software stack for the 21st century warehouse-scale data centres.

Twitter: @CamSysAtScale


Blog: speeding up graph analytics with 10G networks

8 July 2015

Frank McSherry visited CamSaS and we wrote a blog post, "The impact of fast networks on graph analytics", together. The post shows that graph algorithms such as PageRank can benefit hugely from 10G networks, and demonstrates a 20x speedup over a GraphX/Spark stack.

Blog: applications of QJump's guaranteed latency

12 May 2015

Our first blog post, "QJump, or: how to be un-British in the data centre" discusses why the guaranteed latency communication offered by QJump is exciting, and what new applications it may enable.

QJump wins Best Paper award!

4 May 2015

Our QJump paper won a Best Paper Award at NSDI 2015.

Musketeer at EuroSys 2015

21 April 2015

We are presenting our Musketeer paper at EuroSys 2015 on April 22. Find us at the conference, or follow Kermit!

Paper accepted at HotOS 2015

16 March 2015

A paper co-authored by two members of CamSaS has been accepted for publication at HotOS 2015. The paper shows how big data analytics jobs' runtime can be improved by up to 40% by replacing garbage collection (GC) with region-based memory allocation.