Middleware Policies for Intrusion Tolerance Franklin Webber, Partha Pal, Chris Jones, Michael Atighetchi, and Paul Rubel BBN Technologies QuO
Outline Using middleware for defense against intrusions Defense mechanisms Parameterizing defense policies
A Distributed Military Application
A Cyber-Attack Hacked!
An Abstract View Data User Data Source Data Processing (Fusion, Analysis, Storage, Forwarding, etc.) Attacker
Traditional Security Application Attacker Trusted OSs and Network Private Resources Limited Sharing Private Resources
Most OSs and Networks In Common Use Are Untrustworthy Application Attacker OSs and Network Private Resources Limited Sharing Private Resources
Cryptographic Techniques Can Block (Most) Direct Access to Application Attacker OSs and Network OSs and Network Private Resources Limited Sharing Private Resources
Firewalls Block Some Attacks; Intrusion Detectors Notice Others y p t o Application Attacker OSs and Network IDSs Firewalls Raw Resources CPU, bandwidth, files...
Defense-Enabled Application Competes With Attacker for Control of Resources C r y p t o Attacker Application Middleware for QoS and Resource Management OSs and Network IDSs Firewalls Raw Resources CPU, bandwidth, files...
QuO QuO Adaptive Middleware Technology QuO is BBN-developed middleware that provides: interfaces to property managers, each of which monitors and controls an aspect of the Quality of Service (QoS) offered by an application; specifications of the application’s normal and alternate operating conditions and how QoS should depend on these conditions. QuO has integrated managers for several properties: dependability communication bandwidth real-time processing (using TAO from UC Irvine/WUStL) security (using OODTE access control from NAI) QuO
QuO adds specification, measurement, and adaptation into the distributed object model CLIENT Network operation() in args out args + return value IDL STUBS SKELETON OBJECT ADAPTER ORB IIOP (SERVANT) OBJ REF Application Developer CORBA DOC MODEL Mechanism Developer CLIENT Delegate Contract SysCond Network MECHANISM/PROPERTY MANAGER operation() in args out args + return value IDL STUBS SKELETON OBJECT ADAPTER ORB IIOP (SERVANT) OBJ REF Application Developer QuO Developer QUO/CORBA DOC MODEL Mechanism Developer 7
The QuO Toolkit Supports Building Adaptive Apps or Adding Adaptation to Existing Apps QoS Adaptivity Specification QuO Code Generator CORBA IDL Middleware for QoS and Resource Management
Implementing Defenses in Middleware for simplicity: QoS concerns separated from functionality of application. Better software engineering. for practicality: Requiring secure, reliable OS and network support is not currently cost-effective. Middleware defenses will augment, not replace, defense mechanisms available in lower system layers. for uniformity: Advanced middleware such as QuO provides a systematic way to integrate defense mechanisms. Middleware can hide peculiarities of different platforms. for reuseability Middleware can support a wide variety of applications.
Security Domains Limit the Damage From A Single Intrusion host host host router host router host host domain hacked domain
Replication Management Can Replace Killed Processes domain host host host router host router host host domain hacked domain application component replicas QuO replica management
Bandwidth Management Can Counter Flooding Between Routers domain host host host router host router host host domain hacked domain QuO bandwidth management RSVP reservation or packet-filtered link
Other Defensive Adaptations Dynamically configure firewalls to block traffic Dynamically configure routers to limit traffic Dynamically change communication ports Dynamically change communication protocols
Defense Strategy Use QuO middleware to coordinate all available defense mechanisms in a coherent strategy. Our best current strategy has two parts: “outrun”: move application component replicas off bad hosts and on to good ones “contain”: quarantine bad hosts and bad LANs by limiting or blocking network traffic from them and, within limits, shutting them down
Policy Issues for ‘Outrunning’ Where should new replicas be placed? Always in new security domain? Always on a new host? Unpredictably? Should number of replicas change under attack? Increase for protection against stealth? Decrease for more rapid response?
Policy Issues for ‘Containment’ Should quarantine be used? Or rely only on self-shutdown based on local sensors? When is a domain, LAN, or host judged bad? Depends on source of warning? Depends on repeated warnings? Depends on combination of warnings? Is agreement necessary before quarantine? Yes: local decisions are easier to spoof No: global decisions are impeded by flooding
Avoiding Self-Denial-of-Service How to prevent attacker from spoofing defense into quarantining all security domains? Limit number or fraction of quarantined domains? Limit rate of quarantining? Allow later reintegration of quarantined domains?
Conclusion The feasibility of adaptive cyber-defense is being explored. Adaptive cyber-defense is naturally implemented in middleware. A strategy for cyber-defense can be parameterized in several ways.