Search

Large-scale, fault-tolerant, distributed systems are the backbone for many critical software services. Since they must execute correctly in a possibly adversarial environment with arbitrary communication delays and failures, the underlying algorithms are intricate. In particular, achieving consistency and data retention relies on intricate consensus (state machine replication) protocols. Ensuring the reliability of implementations of such protocols remains a significant challenge because of the enormous number of exceptional conditions that may arise in production. We propose a methodology and a tool called Netrix for testing such implementations that aims to exploit programmer's knowledge to improve coverage, enables robust bug reproduction, and can be used in regression testing across different versions of an implementation. As evaluation, we apply our tool to a popular proof of stake blockchain protocol, Tendermint, which relies on a Byzantine consensus algorithm, a benign consensus algorithm, Raft, and BFT-Smart. We were able to identify 4 deviations of the Tendermint implementation from the protocol specification and check their absence on an updated implementation. Additionally, we were able to reproduce 4 previously known bugs in Raft., Comment: Update: - Added missing references - Updated sections "Netrix Unit tests" and "Case studies" to better explain the terminology and results - Added additional figures in section "Our approach" to explain the examples - Typos and grammatical errors

Consider a deep neural network (DNN) that is being used to suggest the direction in which an aircraft must turn to avoid a possible collision with an intruder aircraft. Informally, such a network is well-behaved if it asks the own ship to turn right (left) when an intruder approaches from the left (right). Consider another network that takes four inputs -- the cards dealt to the players in a game of contract bridge -- and decides which team can bid game. Loosely speaking, if you exchange the hands of partners (north and south, or east and west), the decision would not change. However, it will change if, say, you exchange north's hand with east. This permutation invariance property, for certain permutations at input and output layers, is central to the correctness and robustness of these networks. This paper proposes a sound, abstraction-based technique to establish permutation invariance in DNNs with ReLU as the activation function. The technique computes an over-approximation of the reachable states, and an under-approximation of the safe states, and propagates this information across the layers, both forward and backward. The novelty of our approach lies in a useful tie-class analysis, that we introduce for forward propagation, and a scalable 2-polytope under-approximation method that escapes the exponential blow-up in the number of regions during backward propagation. An experimental comparison shows the efficiency of our algorithm over that of verifying permutation invariance as a two-safety property (using FFNN verification over two copies of the network)., Comment: There are 31 pages and 2 figures in this document. This paper was submitted to the 23rd International Conference on Verification, Model Checking, and Abstract Interpretation, but was not selected for publication

This report describes the experiences of Collins Avionics & Communications and SRI International in formally specifying and verifying the microcode in a Rockwell proprietary microprocessor, the AAMP-FV, using the PVS verification system. This project built extensively on earlier experiences using PVS to verify the microcode in the AAMP5, a complex, pipelined microprocessor designed for use in avionics displays and global positioning systems. While the AAMP5 experiment demonstrated the technical feasibility of formal verification of microcode, the steep learning curve encountered left unanswered the question of whether it could be performed at reasonable cost. The AAMP-FV project was conducted to determine whether the experience gained on the AAMP5 project could be used to make formal verification of microcode cost effective for safety-critical and high volume devices.

We present a systematic approach to decompose and incrementally build the proof of correctness of pipelined microprocessors. The central idea is to construct the abstraction function by using completion functions, one per unfinished instruction, each of which specifies the effect (on the observables) of completing the instruction. In addition to avoiding the term size and case explosion problem that limits the pure flushing approach, our method helps localize errors, and also handles stages with interactive loops. The technique is illustrated on pipelined and superscalar pipelined implementations of a subset of the DLX architecture. It has also been applied to a processor with out-of-order execution.

This paper describes a bitvectors library that has been developed for PVS. The library defines a bitvector as a function from a subrange of the integers into (0,1). The library provides functions that interpret a bitvector as a natural number, as a 2's complement number, as a vector of logical values and as a 2's complement fraction. The library provides a concatenation operator and an extractor. Shift, extend and rotate operations are also, defined. Fundamental properties of each of these operations have been proved in PVS.

Formal specification combined with mechanical verification is a promising approach for achieving the extremely high levels of assurance required of safety-critical digital systems. However, many questions remain regarding their use in practice: Can these techniques scale up to industrial systems, where are they likely to be useful, and how should industry go about incorporating them into practice? This report discusses a project undertaken to answer some of these questions, the formal verification of the AAMPS microprocessor. This project consisted of formally specifying in the PVS language a rockwell proprietary microprocessor at both the instruction-set and register-transfer levels and using the PVS theorem prover to show that the microcode correctly implemented the instruction-level specification for a representative subset of instructions. Notable aspects of this project include the use of a formal specification language by practicing hardware and software engineers, the integration of traditional inspections with formal specifications, and the use of a mechanical theorem prover to verify a portion of a commercial, pipelined microprocessor that was not explicitly designed for formal verification.

The main goal of the project was two-fold: First, to investigate the feasibility of formally specifying and verifying a complex commercial microprocessor that was not expressly designed for formal verification. Second, to explore effective ways to transfer the technology to an industrial setting. The choice of the AAMP5 satisfied the first goal since the AAMP5 was not designed for formal verification, but to provide a more than threefold performance improvement while remaining object-code-compatible with the earlier AAMP2, which is used in numerous avionics applications, including the Boeing 737, 747, 757, and 767. To satisfy the technology transfer objective, we had to develop a suitable verification methodology and a formal infrastructure to make the technology usable by practicing engineers. This infrastructure includes techniques for decomposing the microcompressor verification problem into a st of verification conditions that the engineers can formulate and strategies to automate the proof of the verification conditions. The development of the infrastructure was one of the key accomplishments of the project. Most of the infrastructure and methodology are general enough to be reused for other microprocessors, certainly in the verification of another member of the AAMP family. This methodology was used to formally specify the entire microarchitecture and more than half of the instruction set and to verify a core set of eleven AAMP5 instructions representative of several instruction classes. However, the methodology and the formal machinery developed are adequate to cover most of the remaining AAMP5 instructions. Although PVS was the vehicle of the experiment, the methodology is applicable to other sufficiently powerful theorem provers.

This presentation describes a project, formal verification of the microcode in the AAMP5 microprocessor, conducted to explore how formal techniques for specification and verification could be introduced into an industrial process. Sponsored by the Systems Validation Branch of NASA Langley and by Collins Commercial Avionics, a division of Rockwell International, it was conducted by Collins and the SRI International Computer Science Laboratory. The project consisted of specifying in the PVS language developed by SRI a portion of a Rockwell proprietary microprocessor, the AAMP5, at both the instruction set and register-transfer levels and using the PVS theorem prover to prove the microcode correct for a representative subset of instructions. While this presentation includes a brief technical overview, its emphasis is on the lessons learned in using PVS for an example of this size and the implications for using formal methods in an industrial setting. The central result of this project was to demonstrate the feasibility of formally specifying a commercial microprocessor and the use of mechanical proofs of correctness to verify microcode. This is particularly significant since the AAMP5 was not designed for formal verification, but to provide a more than three fold performance improvement, by pipelining instruction execution, while remaining object code compatible with the earlier AAMP2. As a consequence, the AAMP5 is one of the most complex microprocessors to which formal methods have been applied. Another key result was the discovery of both actual and seeded errors. Two actual microcode errors were discovered and corrected during development of the formal specification, illustrating the value of simply creating a precise specification. Two seeded errors were systematically uncovered while doing correctness proofs. One of these was an actual error that had been discovered after first fabrication but left in the microcode provided to SRI. The other error was designed to be unlikely to be detected by walkthroughs, testing, or simulation. Several other results emerged during the project, including the ease with which practicing engineers became comfortable with PVS, the need for libraries of general purpose theories, the usefulness of formal specification in revealing errors, the natural fit between formal specification and inspections, the difficulty of selecting the best style of specification for a new problem domain, the high level of assurance provided by proofs of correctness, and the need to engineer proof strategies for reuse.

This report documents the Phase 1 results of an effort aimed at formally verifying a key hardware component, called Scoreboard, of a Fault-Tolerant Parallel Processor (FTPP) being built at Charles Stark Draper Laboratory (CSDL). The Scoreboard is part of the FTPP virtual bus that guarantees reliable communication between processors in the presence of Byzantine faults in the system. The Scoreboard implements a piece of control logic that approves and validates a message before it can be transmitted. The goal of Phase 1 was to lay the foundation of the Scoreboard verification. A formal specification of the functional requirements and a high-level hardware design for the Scoreboard were developed. The hardware design was based on a preliminary Scoreboard design developed at CSDL. A main correctness theorem, from which the functional requirements can be established as corollaries, was proved for the Scoreboard design. The goal of Phase 2 is to verify the final detailed design of Scoreboard. This task is being conducted as part of a NASA-sponsored effort to explore integration of formal methods in the development cycle of current fault-tolerant architectures being built in the aerospace industry.

Presented here is a formal specification and verification of a property of a quadruplicately redundant fault tolerant microprocessor system design. A complete listing of the formal specification of the system and the correctness theorems that are proved are given. The system performs the task of obtaining interactive consistency among the processors using a special instruction on the processors. The design is based on an algorithm proposed by Pease, Shostak, and Lamport. The property verified insures that an execution of the special instruction by the processors correctly accomplishes interactive consistency, providing certain preconditions hold, using a computer aided design verification tool, Spectool, and the theorem prover, Clio. A major contribution of the work is the demonstration of a significant fault tolerant hardware design that is mechanically verified by a theorem prover.

The design and formal verification of a hardware system for a task that is an important component of a fault tolerant computer architecture for flight control systems is presented. The hardware system implements an algorithm for obtaining interactive consistancy (byzantine agreement) among four microprocessors as a special instruction on the processors. The property verified insures that an execution of the special instruction by the processors correctly accomplishes interactive consistency, provided certain preconditions hold. An assumption is made that the processors execute synchronously. For verification, the authors used a computer aided design hardware design verification tool, Spectool, and the theorem prover, Clio. A major contribution of the work is the demonstration of a significant fault tolerant hardware design that is mechanically verified by a theorem prover.

The work done at ORA for NASA-LRC in the design and formal verification of a hardware implementation of a scheme for attaining interactive consistency (byzantine agreement) among four microprocessors is presented in view graph form. The microprocessors used in the design are an updated version of a formally verified 32-bit, instruction-pipelined, RISC processor, MiniCayuga. The 4-processor system, which is designed under the assumption that the clocks of all the processors are synchronized, provides software control over the interactive consistency operation. Interactive consistency computation is supported as an explicit instruction on each of the microprocessors. An identical user program executing on each of the processors decides when and on what data interactive consistency must be performed. This exercise also served as a case study to investigate the effectiveness of reusing the technology which was developed during the MiniCayuga effort for verifying synchronous hardware designs. MiniCayuga was verified using the verification system Clio which was also developed at ORA. To assist in reusing this technology, a computer-aided specification and verification tool was developed. This tool specializes Clio to synchronous hardware designs and significantly reduces the tedium involved in verifying such designs. The tool is presented and how it was used to specify and verify the interactive consistency circuit is described.