STEALTH SOFTWARE TECHNOLOGIES, INC. — Department of Defense SBIR Phase II: HR001121S0007-12
STEALTH SOFTWARE TECHNOLOGIES, INC. — SBIR Phase II award from Department of Defense.
- Amount
- $1,490,321
- Agency
- Department of Defense · Defense Advanced Research Projects Agency
- Program / Phase
- SBIR · Phase II
- Topic
- HR001121S0007-12
- Solicitation
- HR001121S0007.I
- NAICS
- —
- Place of performance
- CA
- Period
- 2022-06-27 → 2025-09-27
Description
Stealth Software Technologies, Inc. ("Stealth" the prime contractor) along with the assistance of university subcontractors George Washington University and Georgetown University proposes to develop a Human Movement Network (HumaNet)-based anonymous communication system that enables message transmission over large geographic distances and preserves information privacy during message transmission. Our project includes two of the original authors of the HumaNets work and is entitled MARATHON: Mobile Adaptive Relay Anti-Totalitarian Human Overlay Network. We propose to develop a HumaNet system with geolocational hierarchy incorporated into its profile. The geolocational hierarchy will potentially improve the profile-based routing for the HumaNet. In addition, we propose two more features for improving resilience in the presence of a malicious adversary: an error-correcting code and biased random walks. The first, error correcting codes, are used to protect against compromised HumaNet nodes seeking to modify a message in transit. In addition, the use of error correcting codes provides a more efficient tradeoff between message duplicates to protect against malicious adversaries, so that it also allows HumaNet users to send messages of large size. The second, biased random walks, seeks to avoid a weakness of profile-based routing which can result in messages falling into a sinkhole. To circumvent this problem, we will implement the concept of biased random walk is implemented on top of HumaNets where nodes with less favorable profiles decide to take messages with a certain (small) probability. The parameter must be carefully chosen so that the blow up in the duplication of messages remains under control. Our proposed solution further aims to maximize confidentiality of message-transmission meta-data, through the use of two cryptographic tools: The first is conditional disclosure of secrets (CDS). An efficient CDS protocol will implement the phone-to-phone exchange in the HumaNet so that if the potential receiver does not have a satisfying geographic profile (i.e. the potential receiver is unlikely to carry the message closer to its destination), they will not be able to receive a message, thus protecting against information leakage to phones with non-satisfying location profiles. The second tool is an onion encryption scheme built upon a public key infrastructure where a geographic location is associated with a public/private key pair. The use of onion encryption reveals only the next destination to intermediate nodes during message routing, while hiding the final destination.