E-Infrastructure Resilience to Climate Change

Project Details

This CASE project is supported by BT

Key Questions

How can we increase the resilience of e-Infrastructure to climate change effects?

Background

Almost everything that we do today is computing and communications driven, using e-Infrastructure. Recent years have shown that climate change has significant inverse effects on e-infrastructure, e.g. a flood in a small city can invalidate computing resources worldwide. The “blast radius” of severe weather events defines how many people and services will be affected by e-infrastructure failure due to events such as storms, floods, lightings.

Aims of the Project

This project aims to explore ways to increase the resilience of e-infrastructure to climate change. This includes studying the blast radius due of severe weather events, and developing networking and computing solutions increasing resilience of e-infrastructure.

Project Description

Online shopping, streaming media, smart homes, commerce and finance, almost everything that we do today is connected, and depends on our e-infrastructure. By next year, 94% of the world’s computing is expected to run in the cloud, and Covid-19 has further increased our dependence on e-infrastructure. However, current e-infrastructure is not resilient. Despite best efforts, repeatedly local weather events, such as floods and lightning storms, have worldwide ramifications.

Research to date has either focused on traditional infrastructure (e.g., electricity grid) or on data centres as the centre of computing.

This research project will explore the resilience of e-infrastructure to climate change. The project will study the effects of severe weather events on e-infrastructure. It particular, it will consider the propagation of failures within the infrastructure due to local weather events, also known as the “blast radius”.

The project will develop of practical algorithms and architectures for minimising weather effects on e-infrastructure. Examples include changes to deployment models of communications infrastructure or development of automated reaction and mitigation mechanisms within network devices.

Some of the questions likely to addressed by the project are:

  • How do severe weather events propagate across e-infrastrcuture?
  • Can we increase e-infrastructure’s resilience to climate change by adopting new e-infrastructure paradigms?
  • Can we use programmable network devices to react and mitigate the propagation of failures within e-infrastructure?
  • Can we proactively mitigate the effects of severe weather events by migrating compute, storage and network in advance?

To take part in this project, a previous experience in programming (e.g., Python, C/C++) is required, as well as knowledge of computing and networking concepts. Knowledge of programmable hardware (FPGA, programmable network devices, P4) is an advantage. The successful candidate is expected to learn and gain during the project expert knowledge in e-infrastructure, programmable hardware, computing and communications.

 

References:

[1] Zilberman, N., Moore, A. W., & Crowcroft, J. A. (2016). From Photons to Big Data Applications: Terminating Terabits. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 2014.0445.

[2] Cisco. (2018). Cisco Global Cloud Index: Forecast and Methodology, 2016–2021.

[3] Nichols, S. (2018, September 4). Thunderstruck: Azure Back in Black(out) after High Voltage causes Flick of the Switch. The Register. https://www.theregister.co.uk/2018/09/04/thunderstruck_azure_backout/

Methods to be used

The research will focus on the development of practical algorithms and architectures for e-infrastructure, such as networking devices or compute nodes within affected areas.

Specialised skills required

A degree in Electrical or Information Engineering, or related fields (CS, physics) with relevant courses is required.

Knowledge in computer networks and computer architecture is required.

Programming experience (e.g., Python, C) is required.

Previous experience with programmable devices (FPGA, programmable switches) is an advantage.

Please contact Noa Zilberman on noa.zilberman@eng.ox.ac.uk if you are interested in this project