Profiles and Multi-Topology Routing in Highly Heterogeneous Ad Hoc Networks

Audun Fosselie Hansen

Tarik Cicic
Paal Engelstad

Audun Fosselie Hansen Poster, Infocom 2006

27.04.2006

Improving Network Providers Business Case

Ad hoc networks should be regarded as an opportunity rather than a threat to current business All network infrastructures should be available for the customers Customers should also cooperate making their private infrastructure available for fellow customers Improving service and Internet availability

Audun Fosselie Hansen Poster, Infocom 2006

27.04.2006

Rethinking Research Challenges

Scalability
Has been the traditional research focus [1,2] Tailored mechanisms for the weakest links only

Heterogeneity
Devices and wireless media technologies will provide a wide heterogeneity with respect to capabilities and properties [3] Some schemes focus on heterogeneity, but on one parameter only [4,5]

We argue for a more holistic approach that covers not only scalability, but also heterogeneity in terms of many different aspects, simultaneously

Profiling the Routing in Ad Hoc Networks

Device profiles for management of heterogeneous ad hoc networking systems
device type, power supply, energy level, wireless interface, offered data rate, supported and desired routing approach, mobility pattern, etc.

DiffServ-like property classes

to reduce the attribute space.

Configuring the device profile

by the user, operator or automatically based on current context

Capability profiling may improve scalability,

by ensuring that the low-capacity components are not used for data forwarding.
4

Using Multiple Topology Routing (MTR)

Utilizes independent logical topologies to compute different paths for different types of traffic i.e., a router/device maintains different routing tables for different purposes.

Explore the benefits of using MTR in ad hoc networks Using multiple topologies for resilient routing as described in [6] Profile-adjusted topologies
Different routing processes Different traffic

Audun Fosselie Hansen Poster, Infocom 2006

27.04.2006

Research Goals and Plans

Explore the benefits of using profiles and multi-topology routing in ad hoc networks Develop a framework for profile specification and dissemination.
Profile granularity Trade-offs with routing complexity Device heterogeneity Scalability, e.g. in terms of the dissemination of profile information

Algorithms and mechanisms for building and maintaining consistent specialpurpose topologies should be developed.
Avoiding routing loops is a goal in itself.

As a first step: Use MT routing to improve the resilience of ad hoc networks using multi-topology routing as described in [6].
6

The following figures will show

How different devices announce profiles and supported routing schemes
All devices should announce this in its neighborhood

How different purpose topologies could be built based on this information

Topology for proactive routing processes Topology for reactive routing processes Topology for High-demanding applications

How Multi-topology routing can improve resilience

Example of backup topologies for the proactive nodes Example of how a node can move packets to a topology where a failed next hop is isolated

Audun Fosselie Hansen Poster, Infocom 2006

27.04.2006

Routing protocols, Profile

Routing Protocols: OLSR, AODV etc. Profiles: 1, ..., 8 (High capacity and stable, , Low capacity and high mobility)

Node A announces its supported routing protocol and profile in its neighborhood

Audun Fosselie Hansen Poster, Infocom 2006

27.04.2006

Routing protocols, Profile

Routing Protocols: OLSR, AODV etc. Profiles: 1, ..., 8 (High capacity and stable, , Low capacity and high mobility)

Node B announces its supported routing protocols and profile in its neighborhood
B

Audun Fosselie Hansen Poster, Infocom 2006

27.04.2006

Routing protocols, Profile

Routing Protocols: OLSR, AODV etc.

Node C announces its supported routing protocols and profile in its neighborhood

Profiles: 1, ..., 8 (High capacity and stable, , Low capacity and high mobility)

Audun Fosselie Hansen Poster, Infocom 2006

27.04.2006

10

Resulting in: Proactive Routing Topology (e.g. OLSR)

Audun Fosselie Hansen Poster, Infocom 2006

27.04.2006

11

Resulting in: Reactive Routing Topology (e.g. AODV)

Audun Fosselie Hansen Poster, Infocom 2006

27.04.2006

12

Resulting in: Topology for high demanding applications

Audun Fosselie Hansen Poster, Infocom 2006

27.04.2006

13

Improving Resilience using Multi-Topology Routing

Fast local rerouting in connectionless networks like IP-based MANETs is difficult due to problems with looping [7] Our approach is to build multiple logical topologies in such a way that as many nodes as possible are isolated in one of the topologies.
An isolated node will not carry transit traffic We call these backup topologies These will typically be represented as additional routing tables.

When a node detect that the next hop for the packet is not available, it moves the packet to a topology where the next hop is isolated All nodes should have the same view of topologies A node that moves packets to another topology must mark the packets to identify the topology for other nodes in the network In [6] we have investigated this approach for fixed IP networks
Very few backup topologies are needed to isolate every component once Path lengths for recovered traffic are acceptable

Audun Fosselie Hansen Poster, Infocom 2006

27.04.2006

14

Backup Topologies for the proactive nodes

Full Topology

Backup Topology 2
Backup Topology 1

Backup Topology 3
Audun Fosselie Hansen Poster, Infocom 2006 27.04.2006
15

Traffic from S to D in the full Topology

S
Audun Fosselie Hansen Poster, Infocom 2006 27.04.2006
16

Node N detects a failure on the next hop towards D

S
Audun Fosselie Hansen Poster, Infocom 2006 27.04.2006
17

Node N moves traffic to a backup topology 2 towards node D D

S
Audun Fosselie Hansen Poster, Infocom 2006 27.04.2006
18

References
1. C. E. Perkins and P. Bhagwat, Highly dynamic destination sequenced distance-vector routing (dsdv) for mobile computers, in Conference on Communications architectures, protocols and applications, 1994, pp. 234244. D. B. Johnson, Routing in ad hoc networks of mobile hosts, in Workshop on Mobile Computing Systems and Applications, 1994. S. Kurkowski, T. Camp, and M. Colagrosso, Manet simulation studies: the incredibles, ACM SIGMOBILE Mobile Computing and Communications Review, vol. 9, no. 4, pp. 50 61, October 2005. C. Ma and Y. Yang, A prioritized battery-aware routing protocol for wireless ad hoc networks, in 8th ACM international symposium on Modeling, analysis and simulation of wireless and mobile systems, 2005, pp. 45 52. C. E. Jones, K. M. Sivalingam, P. Agrawal, and J. C. Chen, A survey of energy efficient network protocols for wireless networks, Wireless Networks, vol. 7, no. 4, pp. 342358, August 2001 2. 3.

4.

5.

6.
7.

A. Kvalbein, A. F. Hansen, T. Cicic, S. Gjessing, and O. Lysne, Fast IP network recovery using multiple routing configurations, in Proceedings of INFOCOM, Apr. 2006.
M. Shand and S. Bryant, IP Fast Reroute Framework , IETF Internet Draft, March 2006

Audun Fosselie Hansen Poster, Infocom 2006

27.04.2006

19