School of Computer Science & Statistics
Supervisor: Prof. Donal O’Mahony
Future Internet Technologies
My name is Sandeep. I am a PhD student working in the Centre for Telecommunications Value Chain Research ( CTVR ) in Trinity College Dublin, looking at future internet technologies.
My supervisor is Prof. Donal O’Mahony, director of CTVR. He commented, “CTVR is about the the future of the internet and the internet is all around us. In the past ten or fifteen years it has really transformed the telephone industry. These days, we all have mobile phones, Android and iPhones and so on. All of that has lead to a huge amount of bandwidth coming into the network from the edges.”
My research is focused on faster and power-efficient network devices that are used at the core of the internet. Let me explain very simply how the internet is organised and how my research fits in.
Imagine that we have a lot of computers with a lot of files in them that users wanted to share among themselves. There is device called a switch which interconnects all of these computers. The way it actually works is that each of these computers gets a unique IP address, which is similar to a mobile number. This is the basic structure of a private network. We have a lot of these networks in geographically organised regions. We need a special device called a router which interconnects these networks and, to make things more complicated, these devices are often installed in a ‘cloud’ on the internet. Internet companies like Google make users special servers which provide users with features such as email, video streaming, etc. There are also devices like wireless access points, tablets, smart phones which are also constantly making use of the internet services. All of these wireless devices put pressure on the internet infrastructure.
The core of the internet speeds have now increased so that the largest links are operating at speeds of 40G per second up to 100G per second. Right out at the edge, we’re operating with new 10G ethernet interfaces that are now generating packet-loads in the 10 million packets-per-second range. Doing something useful with a packet stream like that typically takes special hardware. The task of my research project is to look at how we can programme that hardware. This throws up many interesting challenges in software engineering and that’s what we’re trying to investigate.