1. New Models and Algorithms for Active Networks Danny Raz and Yuval Shavitt Bell Labs, Lucent Technology www.cs.bell-labs.com/ ~shavitt/actnet.html

2. 2 Outline Motivation Model description Example analysis Summary

3. 3 The Active Bell-Labs Engine An adjunct active engine to any COTS router Only some (mainly) control packets are diverted to the AE Focus on efficient network management MIB router filter Active Engine (AE) manager session 1session 2

4. 4 The Active Bell-Labs Engine (2) Packet Delay depends on whether it passes thru the AE. Processing time in the AE may depend on – data in the packet. –soft state in the AE. MIB router filter Active Engine (AE) manager session 1session 2

5. 5 What is the right model to analyze algorithmic solutions? How to compare the strength of AN architectures? Are active networks efficient?

6. 6 Standard Asynchronous Model Communication is between neighbors A message arrivals triggers computation at a node A single bound on the delay of a communication + computation cycle

7. 7 [Cidon, Gopal, Kutten, 1995] Two bounds on the delay: –C thru the SS –P thru the NCU Caveats: –P is a constant –source routing. Not applicable to datagrams. SS NCU

8. 8 Our Model Two bounds on the delay: –C thru the FF. –P(k) thru the EE. Forwarding is done according to the destination addr. No assumptions on the routing. We use P(k) = P ·k FF Execution Environment (EE) FilterFilter oracle forwarding

9. 9 DARPA Model vs. Our Model NodeOS EE 1 classclass FF Execution Environment (EE) FilterFilter oracle forwarding EE 2 EE 3 IP

10. 10 Performance Measures Communication (Message) complexity - hops traveled by messages Time complexity - time to mission completion. processing complexity - CPU time used.

11. 11 An Application Example: Route Exploration In the model - a node is only aware of its local neighbors. A node wishes to learn the route to some destination. Abstraction of the traceroute program. 45

12. 12 A naïve Solution The source query nodes sequentially. O(n 2 ) messages. O(n 2 C+nP) time.

13. 13 report-en-route A query process advances sequentially. Reports are sent to the source for each query. O(n 2 ) messages. O(nC+nP) time.

14. 14 collect-en-route A query process advances sequentially. Information is collected in the forward direction, and sent by the destination to the source. O(n) messages. O(nC+n 2 P) time.

15. 15 Route Detection Can we do better?

16. 16 Report-every- l Obtain the route length. Initiate collect-en-route in n/l segments of length l. Complexities: –message O(n 2 /l) –time O(nC+l 2 P) For l=n 2/3 : –message O(n 4/3 ); time O(nC+ n 4/3 P)

17. 17 Collect-rec Optimal up to a log factor ! Obtain the route length. Partition the route to two segments. Send results from the second segment using the FF. Perform recursively. Complexities: –message O(n log n); time O(nC+nP)

18. 18 Other Basic Problems Bottleneck detection - computation along a route. Message dissemination to an ad-hoc group. Topology discovery. Computation of a global function.

19. 19 Summary A new model to analyze active network applications. We demonstrated its use to optimize algorithms. Can be used to compare strength of architectures by comparing lower bounds.