UKNOF29 & Internet Society ION Conference

In 2013 HEAnet partially rebuilt and greatly enhanced the resilience of our optical networks in Dublin and around the country. Around the same time we continued our roll out of 100Mbps connectivity to every secondary school in the Republic of Ireland. This talk aims to describe the technology design choices, the implementation process and some of the interesting issues and pitfalls encountered along the way.

Research, design and testing of a licence exempt 5.8GHz wireless Link solution for streaming data back from an offshore marine research buoy.

Agenda

Overview Metro 100G market
Optical Transmission windows
Options
100G LR4
100G DWDM Coherent
100G DWDM Direct detect
Examples
Some test results

Sponsor Presentation

Over the summer BBC R&D tested a few of the technologies we're working on as part of our ip end to end long term research

Featuring content from the Rio World Cup and Commonwealth Games in UHD (4k) with IP/software delivery from the camera to the consumer we expect this to take over from traditional broadcasting.

We'll present how we did it and some of the results

An update on LINX's UK regional peering strategy, where we are today and our plans for the future.  I'll also give some thoughts on what might be achievable in Northern Ireland/Belfast, based on our current experience developing an exchange in Cardiff.

It's been a busy year at UKNOF since our last update. This presentation, prepared for the UKNOF AGM, gives a brief overview of UKNOF's operations, visibility into its governance and finances, what we have achieved in 2013, and some strategy and plans for the future.

UK IPv6 Council was formed in April 2014 as a local chapter of IPv6 Forum. It's a non-profit organisation with the aim of bringing together technical and business  community in the UK and share experience with and knowledge of IPv6.
This presentation is aiming at introducing the organisation to the the UKNOF meeting attendees and raise awareness of the UK IPv6 Council initiatives.

DNS DDoS attacks continue, fueled by open DNS proxies. Now they're stressing resolvers and authorities worldwide using pseudo random subdomains.  In June of 2014 there was a 400% increase in this traffic and popular domains continue to be targeted.  Analysis of recent DNS data reveals other interesting details.  For instance, Response Rate Limiting in authorities appears to aggravate attacks.  
This presentation will cover the latest attack data as well as tests of the major resolvers showing the impact of capabilities to mitigate them, ranging from changes in recursive behaviors to filtering traffic at ingress.

Over recent months ISPs and other operators of Recursive DNS services have been reporting their experiences with unusual DNS query patterns both to the community and to their DNS software providers.    

Increasing attention is now being paid to the performance of recursive servers under different circumstances, including the evolution of various techniques to help manage client queries in a way that favours 'good' and minimises the impact of 'bad'.  

This talk looks at some of the problems that can be encountered and then presents for discussion the experimental features that are being tested with BIND currently as well as introducing other ideas that have been considered but not yet coded by ISC.

Active discussion will be sought on all potential mitigation and tuning strategies, whether internal to DNS software or as part of an organisation's network infrastructure.

For some time there have been concerns that new gTLD strings were already being informally used on the Internet. ICANN commissioned a study in 2013 to assess the extent of the problem. This was used by ICANN to develop a risk mitigation strategy. The key assumption behind that strategy was then examined in a followup study. An explanation of what was done and the findings from these studies are discussed in this presentation.

HEAnet has been trying to encourage IPv6 deployment since... oh, since ever. So have lots of the rest of us. But for something that's taken a decade or two of effort, the numbers still don't look very good. What happened?

More importantly, what do we need to change? There's an area of research called Disruption Theory that teaches us how a less well equipped competitor can beat a dominant incumbent - can we use this to change our approach to IPv6?

Subtitle: What happens when you turn off IPv4

When Bytemark decided to build our own datacentre, we wanted to avoid using IPv4 wherever possible. Almost all of our internal systems are already dual-stacked, and all staff have native IPv6 at home as well as IPv6 via openvpn.

The usual issues with IPv6 in a dual-stack environment not withstanding, we found some new issues when we tried to go IPv6-only.