Interconnection Networks: Practice unites with Theory (INPUT)
Lead Research Organisation:
University of Manchester
Department Name: Computer Science
Abstract
Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.
Organisations
Publications
Erickson A
(2017)
An Optimal Single-Path Routing Algorithm in the Datacenter Network DPillar
in IEEE Transactions on Parallel and Distributed Systems
Erickson A
(2017)
The stellar transformation: From interconnection networks to datacenter networks
in Computer Networks
Erickson A
(2017)
The stellar transformation: From interconnection networks to datacenter networks
in Computer Networks
Erickson A
(2017)
Improved routing algorithms in the dual-port datacenter networks HCN and BCN
in Future Generation Computer Systems
Erickson A
(2020)
Relating the bisection width of dual-port, server-centric datacenter networks and the solution of edge isoperimetric problems in graphs
in Journal of Computer and System Sciences
Erickson A
(2015)
Routing Algorithms for Recursively-Defined Data Centre Networks
Kiasari A
(2015)
Algorithms and Architectures for Parallel Processing
Navaridas J
(2015)
SpiNNaker: Enhanced multicast routing
in Parallel Computing
Navaridas J
(2019)
INRFlow: An interconnection networks research flow-level simulation framework
in Journal of Parallel and Distributed Computing
| Description | The high-level objective of the research was to incorporate elements of theory and practice so as to further research on interconnection networks. These objectives have been met and we have contributed significant new results on server-centric data centre networks all of them supported by research papers published in top journals and/or conferences. Our primary contributions involve: optimal routing algorithms in the data centre network DPillar; significantly improved routing algorithms in the data centre networks HCN, BCN, DCell, FiConn, and recursively defined networks in general (in terms of hop-length, fault-tolerance, load-balancing, latency, and throughput); a new generic paradigm for the construction of dual-port data centre networks, namely stellar networks; and a flow-based simulator INRFlow that is geared specifically towards flow-based simulation in data centres *which has been essential to carry out all the evaluation work*. In addition, there has also been some contributions to other related communication architectures such as SpiNNaker's massively parallel architecture to novel, silicon-photonics interconnects and finally to FPGA-based modelling and simulation of interconnection networks. The interaction between two groups of researchers, one theoretical and one more applied, has been massively successful and a long-term collaboration has been established. The unique skill sets of both teams have been absolutely necessary to undertake the research. |
| Exploitation Route | None yet. |
| Sectors | Digital/Communication/Information Technologies (including Software) |
| Description | There is potential for our results to be implemented in server-centric data centre networks when this paradigm moves from the university research domain and becomes more established within the builders of data centres. However, at present the switch-centric paradigm still dominates. INRFlow is currently being used as the core experimental environment within other related research projects and is being extended accordingly. |
| First Year Of Impact | 2016 |
| Sector | Digital/Communication/Information Technologies (including Software) |
| Impact Types | Economic |
| Title | INRFlow |
| Description | A novel, open-source flow-based simulator of interconnection networks. This |
| Type Of Material | Improvements to research infrastructure |
| Year Produced | 2015 |
| Provided To Others? | Yes |
| Impact | This tool is being used as part of other network-related projects in which the group is involved (EU-funded EXANeSt and EuroEXA). Also it is being used by some of our partners in these projects. We also plan to use them for new research projects. |
| URL | https://gitlab.com/ExaNeSt/inrflow |
| Title | INRFlow - update needed |
| Description | Interconnection network research flow-level evaluation framework |
| Type Of Technology | Software |
| Year Produced | 2016 |
| Open Source License? | Yes |
| Impact | . |
| Title | INSEE |
| Description | Interconnection Networks Simulation and Evaluation Environment |
| Type Of Technology | Software |
| Open Source License? | Yes |
| Impact | INSEE is a lightweight tool that can be used for evaluation of interconnection networks. It has been used by several research groups worldwide. |
| URL | https://sourceforge.net/projects/insee/ |