1. Overview of Research in Dependable Computing Systems Lab
Saurabh Bagchi
ECE & CS
Purdue University
Last updated: June 2016

2. What is Dependable Computing?
We need computer systems that we can depend on in the face of
Naturally occurring faults – hardware malfunction, software bugs
Malicious intrusions – insider attack or external adversaries
Dependability: Property that the system continues to provide functionality, within time bounds, despite the above kinds of failures
What do we do at DCSL?
Build dependability mechanisms - in software, at application or middleware level
Introduce them in practical systems
Evaluate them in practical settings, with accelerated error rate

3. We are part of two centers:
Who Are We?
14 graduate students
2 research scientists
We are part of two centers:
CERIAS
GE-PRIAM
Our support comes from a variety of sources
Federal government: NSF, Department of Defense, Department of Energy, NIH
Private organizations: IBM, Northrop Grumman, Lawrence Livermore, Google, AT&T

4. Thrust #1: Distributed Intrusion Tolerant System
Distributed systems subjected to malicious attacks to services, including hitherto unknown attacks
Objective is to tolerate intrusions, not just detect
Different phases: Detection, Diagnosis, Containment, Response.
Solution approach:
Attack graph based modeling of multi-stage attacks
Algorithms for containment and survivability computation and semi-optimal response selection
Provides situational awareness in the face of imperfect observability
Dealing with zero-day attacks
Work supported by: Missile Defense Agency (MDA), Northrop Grumman, NSF

5. Thrust #1: Distributed Intrusion Tolerant System
Hacker S4 S3

6. Thrust #1: Distributed Intrusion Tolerant System

7. Thrust #2: Debugging Large-Scale Distributed Applications
Goal is to provide highly available applications (e.g., web service) in distributed environment
Perform failure prediction
Perform bug localization
Challenges in today’s distributed systems
Large number of entities and large amount of data
Interactions between entities causes error propagation
High throughput or low latency requirements for the application
Work supported by: NSF, AT&T, Lawrence Livermore National Lab (LLNL)

8. Thrust #2: Predictive Reliability Engine for Cellular Networks
ISP
Backbone
eNodeB
MME
S/P-Gateway
IMS Core
Handoff thresholds
Online data
Call pause
Offline data (GPEH)
P-GW VM
Offline Training
Online predictor
ML Classifier
Extreme Cloud Site

9. Thrust #2: Metric-based Bug Localization
Application
Requests rate, transactions, DB reads/writes, etc..
Middleware
Virtual machines and containers statistics
Operating System
CPU, memory, I/O, network statistics
Hardware
CPU performance counters
Tivoli
Operations Manager
How can we use these metrics to localize the root cause of problems?

10. Thrust #2: Metric-based Bug Localization
Look for abnormal time patterns
Pinpoint code regions that are correlated with these abnormal patterns

11. Thrust #2: Debugging at Extreme Scales
Problem Statement:
Applications being run at large scales – large number of processes (such as, Hadoop clusters) and on large amounts of data (such as, computational genomics applications)
Correctness or performance problem does not show up at small scales, but shows up at large scales
Current debugging techniques cannot handle such problems
SCALE
# OF TIMES LOOP EXECUTES
Training runs
Production runs
Accounting for scale makes trends clear, errors at large scales obvious
RUN #
# OF TIMES LOOP EXECUTES
Is there a bug in one of the production runs?
Training runs
Production runs
Current debugging techniques work process by process, are centralized rather than distributed, and are offline.

12. Thrust #3: Dependable Embedded Wireless Network
Embedded wireless networks have fundamental resource constraints and are often deployed in hostile or uncertain environments
Constraints include: Energy, Bandwidth, Untrusted nodes, Disconnected networks
Goal:
Develop middleware that provides a robust platform keeping environment constraints in mind
Provide detection, diagnosis, and isolation functionality
Precision Agriculture
Our sample application areas
Industrial Monitoring
Smart Grids
Smart Buildings
Surveillance

13. Thrust #3: Dependable Embedded Wireless Network
Solution directions:
Record and Replay:
Software for tracing events on the wireless node
Replaying traced events on a lab server and debugging support from replayed trace
Fastest reprogramming protocol to upload a patch to the network while nodes are deployed in the field
Game theoretic security:
Game theoretic formulation of adversary and defender incentives
Model incorporates multiple stakeholders with partially aligned interests
Probabilistically optimal investment of fixed defense budget
Randomization-based security:
Randomize binary so that adversary cannot reverse engineer and inject attacks
Place canaries strategically in control and data regions
Randomize placements of functions and data
Work supported by: NSF, Sandia

14. Interested. Want to learn more. https://engineering. purdue
Interested? Want to learn more? PI: Saurabh Bagchi