The UMBC Cyber Defense Lab presents

Model Validation for DARPA DPRIVE

Ian Blumenfeld
UMBC and Two Six Technologies

(joint work with Eric Bond, William Harrison, Chris Hathhorn, Paul Li, Matthew Torrence, and Jared Ziegler)

12–1 pm ET, Friday 6 May 2022, via WebEx

Commodity hardware description languages (HDLs) like VHDL and Verilog present a challenge from a high assurance point of view because they lack formalized semantics: when a team of hardware engineers produces a circuit design in a commodity HDL and claims that it correctly implements a pseudocode algorithm, on what basis can that claim be evaluated? A formalized model of the circuit design may be painstakingly created (e.g., in the logic of a theorem prover), but how are the accuracy and faithfulness of that model then established? The distance between the widely adopted commodity HDLs and formal models of hardware has been a well-recognized and persistent impediment to driving formal methods into hardware development.

This talk presents a technique developed at Two Six Technologies, called model validation, that formally connects hardware design and its formal model via a functional, high-level synthesis language called ReWire. Model validation introduces a “model” program to bridge the gap between the hardware design and algorithm by establishing 1) the equivalence of the algorithm to the model and 2) the equivalence of the model to the circuit design. Equivalence between the algorithm and the ReWire model is verified with a ReWire semantics formalized in Isabelle. Equivalence between the ReWire model and the circuit design is established by producing binary circuits from each (using commodity synthesis tools and the ReWire compiler rwc) and then applying an automated binary equivalence checker.

This talk describes our experience applying model validation as part of the DARPA Data Protection in Virtual Environments (DPRIVE) program. DPRIVE aims to develop a novel hardware accelerator to ease computational challenges preventing widespread use of fully homomorphic encryption (FHE) that began with Gentry’s discovery and was improved upon in the PROCEED program. To this end, DPRIVE’s purpose is to design hardware accelerators to improve upon the existing algorithmic gains to FHE. Model validation through ReWire moves formal methods into the practical world, empowering hardware designers to reason about the correctness, safety, and security properties of their designs. In addition, we expect our pipeline to protect hardware supply chains by allowing for a full formal analysis of RTL implementations before tape out.

Ian Blumenfeld is the Research Director for Mathematics at Two Six Technologies.  In that role, he is a principal investigator on multiple DARPA programs, spanning the areas of formal methods, modern cryptography, and applied category theory.  Prior to his work at Two Six, Ian was a formal verification engineer at Apple, where he verified cryptographic properties of the iPhone secure enclave processor.  Ian has worked in roles in and around the federal research space for more than a decade, including five years as an applied research mathematician at the National Security Agency.  Ian is currently enrolled as a part-time Ph.D. student in the UMBC computer science department, working with Dr. Alan Sherman and Dr. Don Engel. Email: itblumenfeld@umbc.edu

Host: Alan T. Sherman, sherman@umbc.edu. Support for this event was provided in part by the National Science Foundation under SFS grant DGE-1753681. The UMBC Cyber Defense Lab meets biweekly Fridays 12-1pm. All meetings are open to the public. Upcoming CDL Meeting: May 13, Enka Blanchard (Digitrust Loria, France)

The UMBC Cyber Defense Lab presents

Model Validation for DARPA DPRIVE

Ian Blumenfeld
UMBC and Two Six Technologies

(joint work with Eric Bond, William Harrison, Chris Hathhorn, Paul Li, Matthew Torrence, and Jared Ziegler)

12–1 pm ET, Friday 6 May 2022, via WebEx

Commodity hardware description languages (HDLs) like VHDL and Verilog present a challenge from a high assurance point of view because they lack formalized semantics: when a team of hardware engineers produces a circuit design in a commodity HDL and claims that it correctly implements a pseudocode algorithm, on what basis can that claim be evaluated? A formalized model of the circuit design may be painstakingly created (e.g., in the logic of a theorem prover), but how are the accuracy and faithfulness of that model then established? The distance between the widely adopted commodity HDLs and formal models of hardware has been a well-recognized and persistent impediment to driving formal methods into hardware development.

This talk presents a technique developed at Two Six Technologies, called model validation, that formally connects hardware design and its formal model via a functional, high-level synthesis language called ReWire. Model validation introduces a “model” program to bridge the gap between the hardware design and algorithm by establishing 1) the equivalence of the algorithm to the model and 2) the equivalence of the model to the circuit design. Equivalence between the algorithm and the ReWire model is verified with a ReWire semantics formalized in Isabelle. Equivalence between the ReWire model and the circuit design is established by producing binary circuits from each (using commodity synthesis tools and the ReWire compiler rwc) and then applying an automated binary equivalence checker.

This talk describes our experience applying model validation as part of the DARPA Data Protection in Virtual Environments (DPRIVE) program. DPRIVE aims to develop a novel hardware accelerator to ease computational challenges preventing widespread use of fully homomorphic encryption (FHE) that began with Gentry’s discovery and was improved upon in the PROCEED program. To this end, DPRIVE’s purpose is to design hardware accelerators to improve upon the existing algorithmic gains to FHE. Model validation through ReWire moves formal methods into the practical world, empowering hardware designers to reason about the correctness, safety, and security properties of their designs. In addition, we expect our pipeline to protect hardware supply chains by allowing for a full formal analysis of RTL implementations before tape out.

Ian Blumenfeld is the Research Director for Mathematics at Two Six Technologies.  In that role, he is a principal investigator on multiple DARPA programs, spanning the areas of formal methods, modern cryptography, and applied category theory.  Prior to his work at Two Six, Ian was a formal verification engineer at Apple, where he verified cryptographic properties of the iPhone secure enclave processor.  Ian has worked in roles in and around the federal research space for more than a decade, including five years as an applied research mathematician at the National Security Agency.  Ian is currently enrolled as a part-time Ph.D. student in the UMBC computer science department, working with Dr. Alan Sherman and Dr. Don Engel. Email: *protected email*

Host: Alan T. Sherman, *protected email*. Support for this event was provided in part by the National Science Foundation under SFS grant DGE-1753681. The UMBC Cyber Defense Lab meets biweekly Fridays 12-1pm. All meetings are open to the public. Upcoming CDL Meeting: May 13, Enka Blanchard (Digitrust Loria, France)

The UMBC Cyberdawgs will host a 48-hour Jeopardy-style Capture The Flag cybersecurity competition, online and in person for UMBC participants.

DawgCTF 2022 will be a Jeopardy-style CTF, so things are pretty simple. Players register online, either in teams or alone. At 1300 EDT on April 29, 2022, the competition will open. Players will have access to a board of challenges, ranging from easy to nigh-impossible, and scored appropriately. Whoever earns the most points by solving challenges wins! The in-person part of the event will take place in room 206 of the Public Policy Building.

The CTF will begin at 1:00 PM EDT on Friday, April 29, 2022, and end at 1:00 PM EDT on Sunday, May 1, 2022, for a total runtime of 48 hours. The maximum team size is four. There will be an in-person component for UMBC participants only.

At 8:00 PM on Friday, April 29^th (7 hours after the start), we will give out some prizes to UMBC competitors, depending on their current position on the scoreboard at 8:00 PM. Only UMBC competitors will be eligible for these prizes. Don’t worry, they’ll be small prizes, mostly just for fun, and the clout and CTFtime credit will go to the “official” winners at competition closing time (1 pm on Sunday).

See the Dawg CTF 2022 site to register, get more information, and join the Discord server.

The UMBC Cyberdawgs will host a 48-hour Jeopardy-style Capture The Flag cybersecurity competition, online and in person for UMBC participants.

DawgCTF 2022 will be a Jeopardy-style CTF, so things are pretty simple. Players register online, either in teams or alone. At 1300 EDT on April 29, 2022, the competition will open. Players will have access to a board of challenges, ranging from easy to nigh-impossible, and scored appropriately. Whoever earns the most points by solving challenges wins! The in-person part of the event will take place in room 206 of the Public Policy Building.

The CTF will begin at 1:00 PM EDT on Friday, April 29, 2022, and end at 1:00 PM EDT on Sunday, May 1, 2022, for a total runtime of 48 hours. The maximum team size is four. There will be an in-person component for UMBC participants only.

At 8:00 PM on Friday, April 29^th (7 hours after the start), we will give out some prizes to UMBC competitors, depending on their current position on the scoreboard at 8:00 PM. Only UMBC competitors will be eligible for these prizes. Don’t worry, they’ll be small prizes, mostly just for fun, and the clout and CTFtime credit will go to the “official” winners at competition closing time (1 pm on Sunday).

See the Dawg CTF 2022 site to register, get more information, and join the Discord server.

The UMBC Cyber Defense Lab presents

Designing Quantum Resistant Key
Exchange Protocols with CPSA

Dr. Edward Zieglar, CSEE, UMBC

12–1 pm, Friday, 15 April 2022
online via WebEx

With developments in quantum computers and algorithms, the public-key systems that we rely upon for secure network communication will become vulnerable to exploitation. Quantum-resistant key exchange protocols are needed to replace our existing vulnerable protocols. Much of the work has focused on developing new mathematical problems that are conjectured to be quantum-resistant as replacements for our current public-key algorithms. We took a different approach, looking to an old secret-key agreement protocol developed by Leighton and Micali at MIT for the Clipper Chip symmetric encryption system. We will present our analysis of the Leighton-Micali key agreement protocol, weaknesses we uncovered with the Cryptographic Protocol Shapes Analyzer (CPSA), and verification of a new protocol based on their ideas that corrects deficiencies in the original protocol.

Dr. Zieglar is an expert in protocol analysis and computer security at the National Security Agency. He is an adjunct faculty member at UMBC and a member of the UMBC Protocol Analysis Lab. Dr. Zieglar earned his Ph.D. in computer science from UMBC working under Dr. Sidhu. Email: eziegl1@umbc.edu

Host: Alan T. Sherman, sherman@umbc.edu. Support for this event was provided in part by the National Science Foundation under SFS grant DGE-1753681. The UMBC Cyber Defense Lab meets biweekly Fridays 12-1 pm. All meetings are open to the public. Upcoming CDL Meetings: April 29, Ian Blumenfeld (UMBC), May 13, Enka Blanchard (Digitrust Loria, France).

The UMBC Cyber Defense Lab presents

Designing Quantum Resistant Key
Exchange Protocols with CPSA

Dr. Edward Zieglar, CSEE, UMBC

12–1 pm, Friday, 15 April 2022
online via WebEx

With developments in quantum computers and algorithms, the public-key systems that we rely upon for secure network communication will become vulnerable to exploitation. Quantum-resistant key exchange protocols are needed to replace our existing vulnerable protocols. Much of the work has focused on developing new mathematical problems that are conjectured to be quantum-resistant as replacements for our current public-key algorithms. We took a different approach, looking to an old secret-key agreement protocol developed by Leighton and Micali at MIT for the Clipper Chip symmetric encryption system. We will present our analysis of the Leighton-Micali key agreement protocol, weaknesses we uncovered with the Cryptographic Protocol Shapes Analyzer (CPSA), and verification of a new protocol based on their ideas that corrects deficiencies in the original protocol.

Dr. Zieglar is an expert in protocol analysis and computer security at the National Security Agency. He is an adjunct faculty member at UMBC and a member of the UMBC Protocol Analysis Lab. Dr. Zieglar earned his Ph.D. in computer science from UMBC working under Dr. Sidhu. Email: *protected email*

Host: Alan T. Sherman, *protected email* Support for this event was provided in part by the National Science Foundation under SFS grant DGE-1753681. The UMBC Cyber Defense Lab meets biweekly Fridays 12-1 pm. All meetings are open to the public. Upcoming CDL Meetings: April 29, Ian Blumenfeld (UMBC), May 13, Enka Blanchard (Digitrust Loria, France).

The UMBC Cyber Defense Lab presents

A Formal Methods Analysis of the Session Binding Proxy Protocol

Kirellos N. Abou Elsaad, UMBC

12-1 pm, Friday, 1 April 2022, via WebEx

Proposed by Burgers, Verdult, and Eekelen in 2013, the Session Binding Proxy (SBP) protocol intends to prevent session hijacking by binding the application session to the underlying network session (i.e., binding the session token to the SSL/TLS shared key). We present a formal methods analysis of SBP using the Cryptographic Protocol Shapes Analyzer (CPSA). Our analysis reveals that SBP relies critically on the successful establishment of a secure SSL/TLS channel, which can be undermined using well-known attacks. Also, we find that SBP allows for the partial hijacking of a session using a tailgating attack. In this attack, the adversary uses the server to inject and execute malicious code inside the client’s browser to extract the session token and forge a valid state-changing request to the server. This attack is not neutralized by SBP because the request contains a valid session token and is sent over the client’s existing SSL/TLS channel.

Kirellos N. Abou Elsaad is a master’s student in computer science at UMBC working under Dr. Sherman and a member of the Protocol Analysis Lab (PAL). email: abou3@umbc.edu

Host: Alan T. Sherman, sherman@umbc.edu. Support for this event was provided in part by the National Science Foundation under SFS grant DGE-1753681. The UMBC Cyber Defense Lab meets biweekly Fridays 12-1 pm. All meetings are open to the public. Upcoming CDL Meetings: April 15, Edward Zieglar (NSA); April 29, Ian Blumenfeld (UMBC); May 13, Enka Blanchard (Digitrust Loria, France).

The UMBC Cyber Defense Lab presents

A Formal Methods Analysis of the Session Binding Proxy Protocol

Kirellos N. Abou Elsaad, UMBC

12-1 pm, Friday, 1 April 2022, via WebEx

Proposed by Burgers, Verdult, and Eekelen in 2013, the Session Binding Proxy (SBP) protocol intends to prevent session hijacking by binding the application session to the underlying network session (i.e., binding the session token to the SSL/TLS shared key). We present a formal methods analysis of SBP using the Cryptographic Protocol Shapes Analyzer (CPSA). Our analysis reveals that SBP relies critically on the successful establishment of a secure SSL/TLS channel, which can be undermined using well-known attacks. Also, we find that SBP allows for the partial hijacking of a session using a tailgating attack. In this attack, the adversary uses the server to inject and execute malicious code inside the client’s browser to extract the session token and forge a valid state-changing request to the server. This attack is not neutralized by SBP because the request contains a valid session token and is sent over the client’s existing SSL/TLS channel.

Kirellos N. Abou Elsaad is a master’s student in computer science at UMBC working under Dr. Sherman and a member of the Protocol Analysis Lab (PAL). email: *protected email*

Host: Alan T. Sherman, *protected email* Support for this event was provided in part by the National Science Foundation under SFS grant DGE-1753681. The UMBC Cyber Defense Lab meets biweekly Fridays 12-1 pm. All meetings are open to the public. Upcoming CDL Meetings: April 15, Edward Zieglar (NSA); April 29, Ian Blumenfeld (UMBC); May 13, Enka Blanchard (Digitrust Loria, France).