Upload
niels-groeneveld
View
186
Download
3
Tags:
Embed Size (px)
Citation preview
Security in Delay Tolerant Networks
Dr. Milena RadenkovicPhD (Nottingham, UK), Dipl. Ing., MSc (Nis, Serbia)
Cyber DefenceNational Security in a Borderless World
17th & 18th May 2010, Swissôtel Tallinn, Estonia
University of Nottingham Cyber Defence, Estonia, 2010
Research Overview
Research interests centre on self-organised networkarchitectures that support interactive multiuser applicationsin unstable and heterogeneous environments
Particular concern is with the design and deployment ofnovel more reliable mobile ad-hoc, delay tolerant networksfor data store and query and routing protocols
Multiple projects in location based pervasive gaming,wearable medical and veterinary applications and massenvironmental monitoring: A Novel Routing Protocol for Large Scale Disconnected (PI, EPSRC), Developing Advanced Collaborative Environments for Life Science
Community (PI, EPSRC) Participate (CI, EPSRC), myGrid (CI, EPSRC) IPERG, (WPLead, EU), MIAS (WPLead, EPSRC)
University of Nottingham Cyber Defence, Estonia, 2010
Disconnection Tolerant Networks
Type of challenged networks wherecommunication opportunities are based onsporadic and intermittent contacts, long disconnections and re-connections may frequently
occur, and the assumption on the existence of an end-to-end path
between the source and the destination is dropped
DTN network features pose fundamentalchallenges to the mechanisms needed to secureDTNs and heavily constrain available securitysolutions
University of Nottingham Cyber Defence, Estonia, 2010
The Internet Architecture
Traditional Internet (wired, wireless, mobile, ad hocnetworks) makes strong assumptions about connectivitysuch as: available end to end paths, low RTTs, high availability to naming,
security services, caching and searching infrastructures to providelocator-based access such as DNS
Wired LAN
4
Cell tower Wireless LAN
MANET
University of Nottingham Cyber Defence, Estonia, 2010
Non Internet-like Architecture
When the connectivity assumptions do not hold Applications break / communication disabled Need for fundamentally new communication and security paradigms Support interoperability among radically heterogeneous networks
and achieve good performance with very large delays andunpredictable loss of connectivity
DTNs use “store, carry and forward” to transfer data and securitymessages
Alan Bob Charlie
University of Nottingham Cyber Defence, Estonia, 2010
Applications of DTNs
Started off as inter-planetary communication but now used for enablingcommunication when the infrastructure is difficult to deploy, expensive to deploy or
available, but a DTN can still improve performance
Military, Interplanetary Disconnected kiosks in rural areas Disaster struck areas Remote sensing applications
But also: Bulk data distribution in urban areas Sharing of individual contents in urban areas Mobile location-aware sensing applications Social Mobile Applications
University of Nottingham Cyber Defence, Estonia, 2010
DTN Examples 1:Inter-Planet Satellite Communication Network Internet Service in Space
(Initial concept of DTN) Characteristics
High Intermittent Connectivity Extremely Long Propagation -
Delay: finite speed of light Low Transmission Reliability:
positioning inaccuracy, limitedvisibility
Low Asymmetric Data Rate Current Projects
InterPlaNetary Internet DARPA JPL MITRE, USC, UCLA,
CalTech, etc.
Security: CCSDS protocol space End to end security space end to end reliability
More recently security DTNprotocol
University of Nottingham Cyber Defence, Estonia, 2010
DTN Examples 2:Military Battlefield Network
No consistent networkinfrastructure and frequentdisruptions
Characteristics High Intermittent Connectivity Mobility, destruction, noise, attack,
interference Low Transmission Reliability:
positioning inaccuracy, limitedvisibility
Low data rate Current Projects
DTN Project @ DARPA Security:
Mainly MANET Security Distribution of CAs in mobile ad hoc
networks cannot provide militarylevel security
Combining a self-organizedapproach with an off-line thirdtrusted parties (TTP) promising
University of Nottingham Cyber Defence, Estonia, 2010
DTN Examples 3:Remote Area Networks
Providing Internet connectivityto rural/developing areas e.ge-mail
Characteristics Intermittent Connectivity Mobility, sparse deployment High Propagation Delay Asymmetric Data Rate: heterogeneous
Current Projects First Mile Solutions, KioskNet
Security requirements: integrity of KioskNet components
(gateways, ferries, kiosk controllersand proxies),
security of kiosk terminals,confidentiality and integrity of userdata despite using untrusted ferries
Security used: standard cryptographic techniques
such as PKI and a transparentencrypted file systems.
University of Nottingham Cyber Defence, Estonia, 2010
DTN Examples 4:Sparse Mobile Ad Hoc Networks
Intermittent Autonomous(Opportunistic)Communications
Even when infrastructure isavailable, this providescheaper alternative to cellularnets e.g. google from the buswithout 3G
Characteristics Intermittent Connectivity Mobility, sparse deployment Large end to end delay
Current Projects DieselNet, CarTel Participate DOME : Diverse Outdoor Mobile
Environment @ UMass Haggle Project @ European Union
Framework Program Security: DTN or PSN security
Bui
lt-in
Acc
eler
omet
erAni
mal
Are
aN
etw
ork
Mul
tiH
opW
irele
ssC
omm
unic
atio
n
Not
ifica
tions
and
Que
ryin
g
Gat
eway
Inte
rnet
Gat
eway
Gat
eway
Inte
rnal
Sto
rage
and
Pro
cess
ing
University of Nottingham Cyber Defence, Estonia, 2010
DTN Security Goals
Due to the resource-scarcity that characterizes DTNs, theemphasis of DTN security is on protecting the DTNinfrastructure from unauthorized access and use Prevent access by unauthorized applications, Prevent unauthorized applications from asserting control over the
DTN infrastructure, Prevent authorized applications from sending bundles at a rate or
class of service for which they lack permission, Promptly detect and discard bundles that were not sent by
authorized users, (early detection within infrastructure rather thanat destination),
Promptly detect and discard bundles whose headers have beenmodified
Promptly detect and disable compromised entities Secondary emphasis is on providing optional end-to-end
security services to bundle applications
University of Nottingham Cyber Defence, Estonia, 2010
DTN security challenges
Security/reliability No trusted infrastructure No standard AAA, no PKI No available fully distributed security algorithms New and different classed of application traffic
The challenges facing the network securitycommunity for supporting such applications arefundamentally profound.
Traditional security approaches for wired, wirelessand ad hoc mobile networks assume that a fullyconnected path between all end points that wishto communicate must exist for trust building to bepossible
University of Nottingham Cyber Defence, Estonia, 2010
How do DTN environments constrain availabletrust building mechanisms
High round-trip times and disconnections Do not allow frequent distribution of a large number of certificates
and encryption keys end-to-end More scaleable to use user’s keys and credentials at neighboring or
nearby nodes.
Delayed or loss of connectivity to a key or certificate server Multiple certificate authorities/key servers desirable but not
sufficient and certificate revocation not appropriate
Long delays messages may be valid for days or weeks, so message expiration
may not be able to be depended on to rid the network of unwantedmessages as efficiently as in other types of networks.
Constrained bandwidth Need to minimize cost of security in terms of header bits
University of Nottingham Cyber Defence, Estonia, 2010
Traditional PKI not applicable
Traditional symmetric cryptography and PKI-basedapproaches are not suitable for DTNs for twomajor reasons. In a PKI, a user authenticates another user’s public key
using a certificate signed by a certificate authority (CA).In a disconnected network, without online access to anarbitrary receiver’s public key or certificate, sending anencrypted message on the fly is not possible
Also, PKIs implement key revocation based onfrequently updated online certificate revocation lists(CRLs) posted by CAs. In the absence of instant onlineaccess to CRLs, a receiver cannot authenticate asender’s certificate.
University of Nottingham Cyber Defence, Estonia, 2010
Identity Based Cryptography not applicable
Identity Based Cryptography (IBC) schemes wherethe public key of each entity is replaced by itsidentity and associated public formatting policiesare not suitable for the security in DTNs for tworeasons: IBC does not solve the key management problem in
DTNs and It is not scalable because it assumes that a user must
know the public parameters for all the trusted parties.
University of Nottingham Cyber Defence, Estonia, 2010
Mobile Ad hoc Key Management Proposals not applicable
Virtual Certificate Authority – Threshold Cryptographyapproach. Not applicable due to no existance of Trusted Third Party
(TTP) Certificate Chaining based on Pretty Good Privacy (PGP)
Not applicable due to insufficient density of certificate graphs,compromised nodes not isolated
Peer-to-peer key management based on mobility Not applicable due to certificate revocation mechanism
Eliminating all forms of on-line and off-line TTP degradessecurity.
Combining a self-organized approaches with an off-line TTPcould provide adequate security
University of Nottingham Cyber Defence, Estonia, 2010
Based on the “bundle” protocol Mandatory protection of the DTN infrastructure
from unauthorized use - detect illegitimate trafficASAP and drop it immediately Hop-by-hop bundle header integrity Hop-by-hop bundle sender authentication Access Control (only legitimate applications/users with
appropriate permissions may inject bundles) Limited protection against DoS by detecting illegitimate
traffic at its first hop and discarding it immediately
Existing Mandatory DTN Security
University of Nottingham Cyber Defence, Estonia, 2010
Existing Optional DTN Security
Optional protection of application data—destination application provided with security evenwhen a router may be compromised End-to-end bundle integrity End-to-end bundle source and destination authentication Replay detection at destination Support for end-to-end payload confidentiality
Security policy router capabilities for enforcing afiner-granularity of access control
University of Nottingham Cyber Defence, Estonia, 2010
Summary of DTN Security Mechanisms
Bundle Authentication Header iscomputed at every sending bundleagent and checked at everyreceiving bundle agent on everyhop along the way from the sourceto destination
Bundle Agent
Bundle Application
Region
Security Policy Router(may check PSH value)
SourceApplication Node
DestinationApplication Node
BAH
Payload Security Header iscomputed once at the sourcebundle agent, carried unchanged,and checked at the destinationbundle agent (and possibly also atsecurity boundary bundle agents).
BAH BAH BAH
PSH
• Source vs. Sender• Destination vs. Receiver
Sender Receiver/Sender
Receiver/Sender
Receiver/Sender
Receiver
Source BundleAgent may enforceaccess control andReject traffic from aBundle application.
University of Nottingham Cyber Defence, Estonia, 2010
DTN Security – Current Issues and Future Efforts
The current DTN security initiative is based on a pre-sharedsecret and involves no trust dynamics mechanisms Works well against external threats but is not applicable to internal
threats i.e. there is no mechanism in place to revaluate a node’s credentials
if a node gets compromised Most recent efforts of the DTN community are directed
towards extending DTN security bundle protocol with novel flexible and fluid
trust building, negotiation and propagation mechanisms based on behavioral modeling, anomalous behavior across
disconnections and non consensus asynchronous partial trust claimingand resolving
Closely aligned with visions and goals of IETF DTN-RG and DARPA
University of Nottingham Cyber Defence, Estonia, 2010
Towards Self-Organised DTN Security
21Direct ReputationCollection
IndirectReputation
Eigen Trust, CentralityMeasures, MobilityProfile
Input Credentials
Exiting DTN Security
n
Message Integrity
Sender Authentication
Access Control
Self-organised DTN Security
University of Nottingham Cyber Defence, Estonia, 2010
Conclusions
Despite a great deal of emerging research dealing withsecurity of mobile ad hoc and DTN networks, supportingsecure routing in an environment with little or noinfrastructure remains a difficult and fundamentallyprofound problem Current decentralised self organised ad hoc security approaches do
not integrate support for disconnections, and security approaches for disconnected networks do not support
dynamics and self organisation
As DTNs get more widely deployed, it becomes increasinglyimportant that more commercial and government effortsare directed towards extending current DTN security toallow dynamic trust management
University of Nottingham Cyber Defence, Estonia, 2010
Thank you!
Any Questions?