Submitted by richard on
News
Building Distributed Temporal Graphs From Event Streams
Submitted by richard on
Workshop paperSecond Workshop on Advances in Mining Large-Scale Time Dependent Graphs (TD-LSG)This paper describes the Raphtory system which is used to analysis large-scale time-varying graph systems. It can ingest streaming graph information and store the complete graph history. It enables queries to be made over the graphs at different points in that graph's history.
Raphtory: Decentralised Streaming for Temporal Graphs
Submitted by richard on
Workshop paperDemo and Doctoral Workshop DEBS (Distributed and Event Based Systems)This work in progress was accepted as a Demo and at the Doctoral workshop for DEBS (Distributed and Event-Based Systems). It shows the early development of a system that ingests events and can create (and eventually query) a dynamic graph.
Emu: Rapid Prototyping of Networking Services
Submitted by richard on
Conference paperProc. USENIX ATCThis paper describes a C# library that can be used to build networked programs which can compile to several target hardware and software platforms. This greatly eases development and debugging. The system is tested using NetFPGA as a target and performs almost as well as hand tuned code.
Demo: Detecting group formations using Beacon technology
Submitted by richard on
Workshop paperACM MobiSys (Demo)This demo shows how Apple's iBeacon technology can be used to track groups of people who are moving together in a crowd.
Faces in the Clouds: Long-Duration, Multi-User, Cloud-Assisted Video Conferencing
Submitted by richard on
Journal PaperTransactions on Cloud Computing (to appear)This paper is a simulation based study of cloud assisted multi-user video streaming. It is based upon two use cases (one related to video poker the other related to MOOCs). The paper looks at strategies for placing cloud locations to facilitate streaming using Amazon EC2 cloud locations. The paper compares a strategy that dynamically picks new locations for cloud hosts as time goes on. Interestingly this seems to provide little benefit compared with simply having a good initial choice of sites even when users may drop into and out of a cloud chat session over the course of many hours.
On rate limitation mechanisms for TCP throughput: a longitudinal analysis
Submitted by richard on
Journal PaperComputer NetworksThis paper is a considerably expanded version of the INFOCOM paper.
Again it argues that TCP is no longer mainly controlled by loss and congestion but instead by algorithms and settings under the control of the sender or receiver deliberately or accidentally designed to restrict throughput for a variety of reasons (for example limiting video sending to the rate at which the viewer is watching).
It contains extended discussion of the methodology and in particular how flight and RTT data was extracted from passive traces.
Detecting Group Formations using iBeacon Technology
Submitted by richard on
Workshop paperHuman Activity Sensing Corpus and Applications WorkshopThis paper looks at how sensor measurements in mobile phones can be used to determine when people are talking in a group.
FLICK: Developing and Running Application-Specific Network Services
Submitted by richard on
Conference paperUSENIX ATCThis paper describes a system for middleboxes that process application level data -- that is reconstructed TCP flows not packets. The system consists of three parts:
1) A language specific to middleboxes that can quickly express data formats and how to process them but in a "safe" way that allows middleboxes to co-exist on the same physical hardware.
2) An abstraction, the task graph, that breaks middlebox logic into small, parallelisable logical units (tasks) connected by channels through which data flows.
3) A system that allows the compiled code to execute in a performant way.FLICK: Developing and Running Application-Specific Network Services
Submitted by richard on
Conference presentationLocation:USENIX ATCSlides:Self-Tuning Service Provisioning for Decentralized Cloud Applications
Submitted by richard on
Journal PaperTransactions on Networks and Service ManagementThis paper considers auction mechanisms that can allow people to bid for resources in cloud computing environments when resources are constrained.
This talk describes FLICK a system for the application-specific middlebox. It consists of three parts:
1) A domain specific language for the middlebox that allows easy development of typical middlebox functions.
2) An abstraction, the task graph, that allows the breaking of middlebox functions into easily parallelisable work units.
3) The system -- this implements the compiled language, handles TCP connections and memory management.
The whole system is comparable in speed to a specialist implementation.