Janet end-to-end performance initiative
Supporting user communities in making optimal use of the Janet Network for high-performance network applications
Started 1 Mar 2015
£250k to £1m
We will identify, document and share best practices on high performance networking, to raise awareness amongst Janet Network connected communities of the issues and factors affecting end-to-end network performance
There is an increasing desire within research and education communities to use more demanding networking applications, such as real-time media or data streaming, low latency haptic applications, or bulk transfers of research data in the order of terabytes or more.
With increasing use of outsourced services and off-site data centres, where applications live outside the home institution, network performance is also becoming ever more important.
What we're doing
We will identify existing and emerging communities that run or are looking to run more demanding network applications, as well as those that may be unaware of the capabilities of the Janet Network. We want to encourage and facilitate discussion with and between those stakeholders.
Our service will include specialist assistance to communities, identifying requirements, helping to troubleshoot problems, and recommending technical solutions. We will also seek to deploy appropriate tools to assist users in visualising their use of the network.
With the involvement of all stakeholders – user communities, computer service departments, network researchers and ourselves – we aim to raise the ‘high water mark’ for network applications, promoting awareness of what is possible on the Janet Network infrastructure.
We also aim to facilitate well-attended, productive workshops and vibrant mail list discussions, and foster strategic community oversight of the project through a dedicated, expert special interest group (SIG).
Case study: large-scale data transfer from the Diamond Light Source
This case study reports on the approach taken to improve the performance of data-intensive science transfers over the Janet Network between Diamond Light Source and its end-user research sites.
The improvements followed the ESnet ‘Science DMZ’ design pattern. Janet-connected member sites that adopt similar methods will also be able to realise the end-to-end performance benefits if they adopt similar principles, allowing them to benefit fully from the power of the Janet network in cases where previously their data may have been shipped by physical media.
We are looking for contributions and input from all stakeholders, to help us shape our work to meet your needs