 |
|
IP network resilience and robustness
We are interested in developing models for failure modes of large and small IP networks, and exploring modifications to the IP routing and forwarding paradigms that would increase a network's overall robustness. For example, we might explore questions like:
-
How graceful are routing failures in the presence of link instabilities? (Characterised from the perspective of end users and ISP customers.) How does choice of topology (vs type and tuning of routing protocols) affect the severity of outages? What changes will enable IP networks to meet the reliability requirements for real-time, interactive, and emergency services?
-
To what degree does TCP's dynamic interaction with transient network problems exacerbate the end-user's perception of network performance degradation? How might this situation be changed or improved? How should the current model of congestion management be improved?
-
Identify the vulnerabilities in existing IP networks to attacks directed against infrastructure elements such as routers, switches, and servers. Explore whether encrypted 'control planes' would make much difference and could be deployed in such a way that ISPs would find it operationally attractive.
-
What does it mean to secure short range wireless LANs (such as 802.11 or Bluetooth) against snooping or active interference? Users are often the weakest link - can we engineer a plug and play systems-level security solution that requires little user involvelment?
-
Should security be about impenetrable defenses, or detection and reaction? How do we automate the differentiation between legitimate and suspicious network use?
|
|
