Securing Nomads:

1
Securing Nomads The Case For Quarantine,
Examination, Decontamination
Kevin Eustice, Shane Markstrum, V.
Ramakrishna, Dr. Peter Reiher, Dr. Leonard
Kleinrock, Dr. Gerald Popek Laboratory for
Advanced Systems Research UCLA Computer
Science Annual Computer Security Applications
Conference 2003
2
In a Nutshell

• Problem summary
• Networks do little to monitor or control entry
• Exploited or vulnerable nomadic devices freely
  move around
• Other devices may victimize or fall victim to
  these devices
• A proposed model QED
• Quarantine devices upon entrance
• Examine devices as required by environment
• Decontaminate devices to repair or update

Introduction Challenges The Paradigm
Conclusion
3
New Trends In Nomadicity

• Users
• Frequently change networks, taking their devices
  with them
• Carry misconfigured and vulnerable software with
  them from locale to locale
• Pick up electronic hitchhikers (viruses,
  malicious agents, other malcode) from other
  nomads they encounter

Introduction Challenges The Paradigm
Conclusion
4
Scenario nomadic blaster propagation
Local Café
5
Scenario nomadic blaster propagation
Bobs Office
6
Traditional Security Ignores Nomadic Devices

• Wireless focus has been on better
• Authentication
• Encryption
• Wired and wireless devices promiscuously enter
  and leave networks
• Little accountability in existing paradigm
• Reactive security, not proactive

Introduction Challenges The Paradigm
Conclusion
7
Life will only get worse

• Pervasive Computing is coming
• Pervasive paradigm implies many more attack
  vectors and potential attackers
• Abundant confidential and important personal
  information
• Some possibilities
• Trojan horses in consumer electronics
• PDA-carried viruses
• Wireless parasites

Introduction Challenges The Paradigm
Conclusion
8
Characteristics of the Environment

• Many, many affected users and devices
• Heterogeneous OS/application space
• Dynamic, often short-lived network membership
• Mostly benevolent but non-technical users
• Minimal system administration available
• Where do we go from here?

Introduction Challenges The Paradigm
Conclusion
9
QED
Worker
Decontaminate Work with device to repair
vulnerabilities!
Bobs Office
10
Quarantine

• Typically, there are two immediate types of
  desired quarantine
• Isolation from outside world
• Many networks partially do this
• Often imperfectly
• Isolation from peers
• Few networks do this
• Just as important

Introduction Challenges The Paradigm
Conclusion
11
Quarantine

• Some mechanisms to quarantine devices include
• Routing restrictions at gateway
• Voluntary isolation by device
• DENY firewall rules on peers
• MAC address-based forwarding restrictions in
  Access Point
• Quarantine wireless network outside firewall

Introduction Challenges The Paradigm
Conclusion
12
Examination

• Many possible alternatives
• Software package analysis
• Network profiling
• Configuration analysis
• File checksum examination
• Virus scan

Introduction Challenges The Paradigm
Conclusion
13
Decontamination

• Assist device in complying with local policy
• Work with device to fix problems
• Update software packages, configurations
• Ask device to disable certain services while in
  this network, etc.

Introduction Challenges The Paradigm
Conclusion
14
Work in ProgressQED Prototype
Introduction Challenges The Paradigm
Conclusion
15
Scenario QED Prototype design
Default drop rules on Worker nodes have already
isolated them from the untrusted Client.
IPsec tunnel
IPsec tunnel
IPsec tunnel
Authenticated DHCP, w/IPsec key insertion
RPM Examination
Package Update
Security Manager
IPsec tunnel
UCLA CS
16
Open Issues

• Overhead management
• Privacy
• Leveraging trust relationships
• Heterogeneity

Introduction Challenges The Paradigm
Conclusion
17
Big Picture

• QED is a component of Panoply, UCLAs pervasive
  computing project
• We think QED is a step towards more secure
  pervasive environments

Introduction Challenges The Paradigm
Conclusion
18
Conclusions

• Existing security mechanisms are insufficient for
  emerging pervasive computing paradigm
• Security needs to be proactive
• QED is the first system to address these issues

Introduction Challenges The Paradigm
Conclusion
19
References

• For more info
• Contact qed_at_lasr.cs.ucla.edu
• Kevin Eustice, Leonard Kleinrock, Shane
  Markstrum, Gerald Popek, Venkatraman Ramakrishna,
  Peter Reiher . Enabling Secure Ubiquitous
  Interactions . In the proceedings of the 1st
  International Workshop on Middleware for
  Pervasive and Ad-Hoc Computing.
• Kevin Eustice, Leonard Kleinrock, Shane
  Markstrum, Gerald Popek, Venkatraman Ramakrishna,
  Peter Reiher . Wi-Fi Nomads The Case for
  Quarantine, Examination and Decontamination . To
  appear in the proceedings of the New Security
  Paradigms Workshop 2003.