Transcript Transient Network Architecture (TNA)

Transient Network
Architecture
http://hdl.handle.net/2118/tna
Joud Khoury
University of New Mexico, ECE department
CNRI
Handle System Workshop, Washington DC
June 21
Credits
• Funded by NSF Future Internet Design
(FIND) Grant CNS-0626380
• Website: http://hdl.handle.net/2118/tna
• People
 Henry Jerez, CNRI
 Joud Khoury, Chaouki Abdallah, Greg
Heileman, Pramod Jamkhedar, Wennie Shu,
Jorge Crichigno, Jorge Piovesan – UNM
Outline
• Overview of TNA
• Mobility and Persistent Identification
• How/where is it applied?
 InterMesh instantiation of TNA
 VoIP using handles
 Digital Rights Management within TNA
• References
What is TNA
• It is an architecture that postulates that:
 All networks can be reduced to a particular
case of persistently identified, transient,
mobile, abstract entities that group into
particular association
• It is an abstraction based persistent
communication network for transient digital
entities
TNA Principles
• Mobility and Ad Hoc characteristics as
basic requirements
• Abstraction as the basis of internetworking
and functionality
• Persistent Identification
Mobility and Ad Hoc
characteristics
• Mobility is the ability of nodes to change
association without breaking referential integrity.
• All components of the architecture are to be
considered transient and mobile
 The components must operate in both
 Connected = Full cohesive communication with the larger
structure is assumed
 Disconnected = limited or no connectivity with other
groupings is possible
• While consolidation is possible; all processes,
services and mechanisms should assume a
mobile deployment environment.
Persistent Identification
• Persistent Identifier=Handle
• Persistently identify:
 globally Digital entities:
 Network components
 Communicating entities
 Services
 Processes
• Identification is based on a set of unique naming
spaces with a distributed resolution on a need to
know basis
How do we use PI’s
• We identify all network entities with
persistent identifiers
• We use these identifiers to route all traffic
in the network
• We identify particular network associations
with persistent identifiers
• We provide secure distributed
administration
• This enables seamless mobility
How do the pieces fit together
What can it do
• Enables new transmission paradigms
• It can move functionality at will
• Allows current and future network to
coexist and seamlessly integrate
Current Research Tracks
• The MESH Network AoI Instantiation –
InterMesh and voice over Intermesh
• Agent Coordination
• Basic DRM expressiveness
InterMesh Motivation
Various types of local networks; WMNs, WSNs,
PANs.
WMN particularly interesting? Advantages,
connectivity models, suitable for WLAN, WMAN,
WVAN, include ad-hoc, …
A growing need for inter-networking among
heterogeneous networks
IP? Some limitations: overloading the address,
absence of trustworthiness
Fundamental design shifts needed -> InterMesh
Motivation
A growing need for inter-networking among heterogeneous (mesh) networks
L3
Persistent Identification (PI) Layer
Mesh Interworking
L2 – 802.11s
(expected)
Topology learning
Routing/forwarding
Measurement
Medium Access Control
L1
IEEE 802.11 PHY
Management
QoS
Architecture model
Node
Core
AoI-1
AoI-2
AoI: Area of
Influence
Architecture
Nodes
Ref
Core
Pi-1
Pi-2
PI Entity
Neutral
Environment
PI
L2
L1
Agents
AoI-1
Agents
AoI-2
Pi-1
Pi-2
Pi-4
Pi-3
AoI: Area of
Influence
A closer look at local delivery
Pi-2
Local A: ZZ
How does Pi-1 know Pi-2 location?
Address resolution – ARP
Core
Agent 2
AoI-1
Pi-1
Local A: AA
Agent 1
AoI-2
Agent 3
AoI-3
Inter-network
1.
Pi-2
Local A: ZZ
Is Pi-3 in Ao1-1? ARP
2. Pi-3 is not in the local network -> send the data to the Agent2
3. Agent2 routes the data
4. Agent3 sends the data to BB
Internet
Agent 2
Agent 3
…
BB
Pi-1
Local A: AA
Agent 1
…
Mike
Pi-3
…
AoI-1
Local A
…
PI
Agent 3
Pi-3
Local A: BB
AoI-2
AoI-3
Mobility
Pi-2
Local A: ZZ
A proactive discovering protocol to keep bound with the core
Agent 2
Core
Agent 3
AoI-1
Agent 1
Pi-1
Local A: AA
Pi-3
Local A: BB
AoI-2
AoI-3
Previous work – VoIP Sessions and
Mobility
Home Domain: hdomain
SIP Proxy/Registrar
(2118/hproxy)
SIP Proxy/Registrar
(10.200/fproxy)
Foreign Domain: fdomain
Sessions and Mobility
Home Domain: hdomain
SIP Proxy/Registrar
(2118/hproxy)
2
3
Handle System
c
a
1
SIP Proxy/Registrar
correspondent
(c_user)
Foreign Domain: cdomain
R
TE
S
GI
RE
INTERNET
d
TER User:
3 r_user
REGIS
Handle: 2118/r_user
SIP Proxy/Registrar
(10.200/fproxy)
Foreign Domain: fdomain
Traditional traffic flow
Proposed traffic flow
H-SIP Abstraction
• SIP users and Proxy servers identified with handles
instead of URI and Domain names eliminating any
domain binding
Proxy Handle
User Handle
Registration - Measurements
• Average Registration times
10,000 samples dispersed
over 10 days
  39ms
t A  5tc
Call Establishment - Measurements
is the diff in cumulative RT delay
Note: Presumably large geographical separation
between the roaming user and his home server
We outperform as long as
>x
IDEA
Indirect DRM Evaluation Architecture
• Rely on persistent identifiers to convey
Rights information
• All content and Users identified with
Persistent Identifiers. All licenses and
rights identified with persistent identifiers
• Use a dynamic evaluation mechanism that
relies on a layered rights expression and
enforcement model
Layered Model
• Persistent identifiers
allow referential
integrity at every layer
• Layers are not only
logically but physically
independent
Dynamic evaluation
• In TNA where even services are mobile; DRM
evaluation along with validation resources are
based on persistent Identifiers
• Persistent Identifiers weave the different layers
and services together
• DRM computation is therefore a heterogeneous
diverse ecosystem rather than a vertically
integrated solution
TNA principles and Rights
Mobility
• Based on TNA principles all components of the
architecture are mobile and transient ness is
assumed
• The system can then use opportunistic
connectivity and realm based interconnection to
conform new enforcement areas
• The System provides first class presence to all
members and their interests: From the owned to
the final consumer and provides them with the
flexibility they need to operate in a more real
environment.
Current Application models
• Heterogeneous License compatibility and
evaluation
• Intrinsic authorization and validation
• New features and traditional behaviors:
 Loans
 Second level market
 New Business models
References
•
TNA
 H. Jerez, J. Khoury, and Chaouki Abdallah, “The Transient Network Architecture”, in
arXiv.
•
InterMesh instantiation of TNA
 J. Khoury, J. Crichigno, H. Jerez, C. Abdallah, W. Shu, and G. Heileman, “The intermesh
network architecture,” under review IEEE Network Magazine.
•
VoIP using handles
 J. Khoury, H. Jerez, C. Abdallah “Efficient User Controlled Inter-Domain SIP Mobility
Authentication, Registration, and Call Routing”, to appear in 1st International Workshop
on Security and Privacy, SPEUCS 2007, Philadelphia, PA, August 2007.
 J. Khoury, H. Jerez, C. Abdallah “H-SIP: Inter-domain SIP mobility: Design”, in
Consumer Communications and Networking Conference, CCNC 2007, Las Vegas, NV,
Jan 2007.
•
Digital Rights Management within TNA
 G. L. Heileman and P. A. Jamkhedkar, DRM Interoperability Analysis from the
Perspective of a Layered Framework, Proceedings of the 5th ACM workshop on Digital
Rights Management, Nov. 2005, Virginia, USA.
 P. A. Jamkhedkar, G. L. Heileman and Ivan Martinez-Ortiz, The Problem With Rights
Expression Languages, Proceedings of the 6th ACM workshop on Digital Rights
Management, Oct-Nov. 2006, Virginia, USA.