Computer Science and Engineering Department

University of Minnesota
Wireless Networking Seminar

From Ad Hoc to ICEBERG: differences

in two wireless network environments

1
 Outline
 Ad Hoc
 What is ad hoc network?
 Why ad hoc network?
 What are the interesting research topics?
 ICEBERG
 What is ICEBERG?
 Why study ICEBERG?
 How can we take it further?

2
 What are Ad Hoc Networks
 In Latin, ad hoc means "for this," further meaning
"for this purpose only.”
 An ad-hoc network is a LAN or other small network,
especially one with wireless connections, in which
some of the network devices are part of the
network only for the duration of a communications
session or, in the case of mobile or portable devices,
while in some close proximity to the rest of the
network.

3
 Definition for Mobile Ad-hoc
 A "mobile ad hoc network" (MANET) is an
autonomous system of mobile routers (and
associated hosts) connected by wireless links--the
union of which form an arbitrary graph. The routers
are free to move randomly and organize themselves
arbitrarily; thus, the network's wireless topology
may change rapidly and unpredictably. Such a
network may operate in a standalone fashion, or
may be connected to the larger Internet. -----
--- IETF

4
 Characteristics of ad hoc wireless
network
 Autonomous (no infrastructure !);
 Wireless link based; (bandwidth constraint)
 Dynamic topology; (Due to movement or
entering sleep mode);
 Rely on batteries for energy; (Power-
constraint)
 Limited physical security;

5
 Why ad hoc wireless networking?
 Technical side:
 wireless devices need to be connected;
 increased performance/cost ratio on
devices
 Internet compatible standards-based
wireless systems;
 Market side:
 mobile computing; wearable computing;
military applications; disaster recovery;
robot data acquisition

6
 Research Challenges (I)
 MAC layer problems:
 Link layer reliability
 QoS at MAC layer
 Power conservation
 Network layer problems: Mobile IP
 Routing;
 QoS
 Power conserving
 Multicast

7
 Research Challenges (II)
 Transport layer problems: (TCP over Ad hoc)
 End-to-end reliability?
 Congestion control?
 QoS?
 Application layer:
 Security?
 QoS?
 Inter-layer interactions;
 Internetworking with internet;

8
 Main problem: Routing
 Standard (Mobile) IP needs an infrastructure
 Home Agent/Foreign Agent in the fixed network
 DNS, routing etc. are not designed for mobility
 No infrastructure in Ad hoc networks
 Main topic: routing
 no default router available
 every host (node) should be able to forward packets

9
 Routing in an ad-hoc network

N1 N1
N2 N3
N3 N2

N4 N4
N5 N5

time = t1 good link time = t2

weak link

10
 Traditional routing algorithms
 Distance Vector
 periodic exchange of messages with all

physical neighbors that contain information

about who can be reached at what distance
 selection of the shortest path if several

paths available
 Link State
 periodic notification of all routers about the

current state of all physical links

 router get a complete picture of the network

11
 Problems of traditional routing algorithms

 Dynamic of the topology

 frequent changes of connections, connection
quality, participants
 Limited performance of mobile systems
 periodic updates of routing tables need energy
without contributing to the transmission of user
data, sleep modes difficult to realize
 limited bandwidth of the system is reduced even
more due to the exchange of routing information
 Problem: protocols have been designed for fixed
networks with infrequent changes and typically
assume symmetric links

12
 Routing (Unicast)
 Table Driven: DSDV, WRP, etc
 On-demand Driven: AODV, TORA, DSR, ABR,
SSR, ……
 Zone Routing Protocol (ZRP)

13
 DSDV (Destination Sequenced Distance
Vector)
 Expansion of distance vector routing
 Sequence numbers for all routing updates
 assures in-order execution of all updates
 avoids loops and inconsistencies
 Decrease of update frequency
 store time between first and best announcement of a path
 inhibit update if it seems to be unstable (based on the
stored time values)

14
 Dynamic source routing (DSR)
 Split routing into discovering a path and maintaining
a path
 Discover a path
 only if a path for sending packets to a certain destination
is needed and no path is currently available
 Maintaining a path
 only while the path is in use one has to make sure that it
can be used continuously

 No periodic updates needed!

15
 Dynamic Source Routing
– Internet-Draft
 Characteristics:
 On-demand
 Unidirectional links and asymmetric routes are supported
 Route Discovery:
 S-D route is included in the header of each packet.
 Nodes forwarding or overhearing data packets may cache
multiple routes for any D for future use (uni-directional?)
 Route Maintenance: on-demand
 Link failure detection: MAC layer (802.11) or Passive ACK
or clear request for ACK
 Link ERR is propagated to source
 Use an cached new route or rediscover
16
 Dynamic Load-Aware Routing
 On-demand, backward learning
 S floods REQ, D choose route by
-- Total buffered packets, Average buffered packets, or
Least number of congested routers
 D detects over-loaded route dynamically and initiates
route-setup procedure to S.
 Load information in I is piggybacked periodically on data packets
 When link failure, the upstream I sends ERR to S and
removes its entry. S initiates new route setup procedure.
 I does not reply REP even it knows a route to D

17
 Mitigating routing misbehavior
 It is impossible to build a perfect network
 Routing denial of service
 Unexpected events, bugs, etc.
 Incorporate tools within the network to detect and
report on misbehavior
 Route only through trusted nodes
 Requires a trust relationship
 Requires key distribution
 Trusted nodes may still be overloaded or broken or compromised
 Untrusted nodes might perform well
 Detect and isolate misbehaving nodes
 Watchdog detects the nodes
 Pathrater avoids routing packets through these nodes
18
 Routing (Multicast)
 Multicast is still a hot topic even in Internet;
 In Ad Hoc, besides of those problems in
traditional Internet, such as congestion
control, routing for multicast is another big
problem;

19
 Other researches on Routing
 QoS support routing;
 Power conserving routing;

20
 ICEBERG
 http://iceberg.cs.berkeley.edu/
 ICEBERG: Internet-based core for CEllular networks
BEyond the thiRd Generation
 Internet-based integration of telephony and data
services spanning diverse access networks
 Leverage Internet’s low cost of entry for service creation,
provision, deployment and integration

21
 Why ICEBERG
 3G+ will enable many communication devices and
networks – diversity
 Mobility for transparent information access
 New applications: audio, video, multimedia

22
 Design Goals
 Potentially Any Network Services (PANS):
 Network and device independent
 Personal Mobility:
 person as communication endpoint; requires a single
identity for an individual - iUID
 Service Mobility:
 seamless mobility across different devices in the middle of
a service session
 Easy Service Creation and Customization
 Scalability, Availability and Fault Tolerance
 Operation in the Wide Area
 Security, Authentication and Privacy 23
 ICEBERG Architecture Overview
Access Network
Plane PSTN GSM Pager

IAP IAP IAP IAP IAP

ICEBERG SF iPOP A NY iPOP

Network
Plane IAP
CA
SF iPOP B NY iPOP PR
PAC
APC
Clearing House NMS

ISP Plane
ISP1 ISP2 ISP3
24
 ICEBERG Components
 ICEBERG Access Point (IAP):
 A gateway serves as a bridge
 Call Agent (CA):
 call setup and control
 Name Mapping Service (NMS):
 mapping between communication endpoint and the iUID
 Preference Registry (PR):
 stores user profile
 Personal Activity Coordinator (PAC):
 tracks dynamic info of a person that is of her interest
 Automatic Path Creation Service (APC):
 establishes and manipulates data flow 25
 iPOP on Cluster Computing
Platforms
 Ninja Base and Active Service Platform (AS1)
 Clusters of commodity PCs interconnected by a
high-speed SAN, acting as a single L-S computer
 mask away cluster management problems
 Load balancing, availability, failure management
 Ninja: highly available service initiation
 Redirector stub
 Good for long running services such as web servers
 AS1: fault tolerant service session
 Client heartbeat with session state
 Good for session-based services such as video conferencing
26
 An Illustration

Alice 3 5 Bob
Call 4 Call
Agent Agent 7
1 2 8
PR NMS PR NMS IAP
IAP

Clearing House

27
 What’s their difference?
 ICEBERG is an integrated service architecture
to link any digital network with the Internet.
 Ad Hoc is in the wireless network domain.
 Put them together, some amazing application
may be available.

28