Interactive Editing Techniques for Subsetting and Dialecting Network Protocols
Abstract
Interactive Editing Techniques for Subsetting and Dialecting Network Protocols We propose to develop techniques and tools for interactively editing built-in network protocol stacks. The proposed editing techniques aim to address problems of bloated software and homogeneous ecosystems of standard network protocols that have larger attack surfaces than necessary. Inparticular, our approach allows for debloating (i.e., eliminating features) and diversifying (i.e., breaking homogeneity) in the context of network protocol stacks; that is, it allows for subsetting and dialecting protocol implementations. The key innovation of the proposed techniques is to make the editing techniques interactive and to provide an empirical verification phase, when applying themodifications to the original protocol stack. In this proposed work, we will develop an end-to-end system, called PROTOSS, which enablesprotocol subsetting and dialecting features on top of its core techniques, namely, interactive editing and verification. At a high level, PROTOSS works as follows. Given a set of client and server programs, it runs all the provided programs within a concolic execution environment with well described, common inputs that network administrators want to support. During the concolic execution, PROTOSS represents program logic and interactions in a symbolic form by using a human readable syntax and annotations of each network communication. Using the symbolic representation of known communications, the network administrators (or programmers) perform subsetting and/ordialecting of the protocols by either modifying the extracted symbolic representation or annotating internal data structures. Whenever such an interaction happens, PROTOSS first estimates the impact (e.g., potential harmfulness) of the user-provided modifications. Then, it synthesizes the newbinaries that reflect these modifications, possibly after additional interactions with the user based on the results of impact analysis. Finally, it empirically verifies the modified protocol using both existing and newly generated inputs.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jul 27, 2018
- Source ID
- N000141812662
Entities
People
- Taesoo Kim
Organizations
- Georgia Tech Research Corporation
- Office of Naval Research
- United States Navy