Abstractions for Practical Virtual Machine Replay

Burtsev, Anton; Johnson, David E.; Hibler, Mike; Eide, Eric; Regehr, John

doi:10.1145/2892242.2892257

Cited by 10 publications

(6 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…VM-level: VM-level hypervisors like XenTT [10], Re-Virt [14,15], ReTrace [45], and TTVM [23] can time-travel entire virtual machines, but at the expense of large program traces and slow time-travel. Reverse-step debugging, like that provided by MCFLY, requires time-travel at interactive speeds to be practical.…”

Section: Time-travel Debuggingmentioning

confidence: 99%

McFly: Time-Travel Debugging for the Web

Vilk,

Berger,

Mickens

et al. 2018

Preprint

View full text Add to dashboard Cite

Time-traveling debuggers offer the promise of simplifying debugging by letting developers freely step forwards and backwards through a program's execution. However, web applications present multiple challenges that make timetravel debugging especially difficult. A time-traveling debugger for web applications must accurately reproduce all network interactions, asynchronous events, and visual states observed during the original execution, both while stepping forwards and backwards. This must all be done in the context of a complex and highly multithreaded browser runtime. At the same time, to be practical, a time-traveling debugger must maintain interactive speeds.This paper presents MCFLY, the first time-traveling debugger for web applications. MCFLY departs from previous approaches by operating on a high-level representation of the browser's internal state. This approach lets MCFLY provide accurate time-travel debugging-maintaining JavaScript and visual state in sync at all times-at interactive speeds. MCFLY's architecture is browser-agnostic, building on web standards supported by all major browsers. We have implemented MCFLY as an extension to the Microsoft Edge web browser, and core parts of MCFLY have been integrated into a time-traveling debugger product from Microsoft.

show abstract

Section: Time-travel Debuggingmentioning

confidence: 99%

McFly: Time-Travel Debugging for the Web

Vilk,

Berger,

Mickens

et al. 2018

Preprint

View full text Add to dashboard Cite

show abstract

“…As in previous work [20,35,13] we measure application progress using CPU hardware performance counters. Ideally we would count retired instructions leading up to an asynchronous event during recording, and dur-ing replay program the CPU to fire an interrupt after that many instructions have been retired -but this approach needs modifications to work in practice.…”

Section: Asynchronous Eventsmentioning

confidence: 99%

“…Other work [20,13] avoids those uniqueness issues by counting the number of branches retired; an instruction at a given address can't be reached more than once without an intervening branch (except for repeating string instructions, which require the ECX/RCX register to be added to the program-state identifier). Unfortunately the branch counter is nondeterministic; for example a transition out a hardware interrupt is counted against the user process, but invisible to RR.…”

Section: Alternative Approaches To Nondeterministic Countersmentioning

confidence: 99%

See 1 more Smart Citation

Engineering Record And Replay For Deployability: Extended Technical Report

O'Callahan,

Jones,

Froyd

et al. 2017

Preprint

View full text Add to dashboard Cite

The ability to record and replay program executions with low overhead enables many applications, such as reverse-execution debugging, debugging of hard-toreproduce test failures, and "black box" forensic analysis of failures in deployed systems. Existing record-andreplay approaches limit deployability by recording an entire virtual machine (heavyweight), modifying the OS kernel (adding deployment and maintenance costs), requiring pervasive code instrumentation (imposing significant performance and complexity overhead), or modifying compilers and runtime systems (limiting generality). We investigated whether it is possible to build a practical record-and-replay system avoiding all these issues. The answer turns out to be yes -if the CPU and operating system meet certain non-obvious constraints. Fortunately modern Intel CPUs, Linux kernels and user-space frameworks do meet these constraints, although this has only become true recently. With some novel optimizations, our system RR records and replays real-world lowparallelism workloads with low overhead, with an entirely user-space implementation, using stock hardware, compilers, runtimes and operating systems. RR forms the basis of an open-source reverse-execution debugger seeing significant use in practice. We present the design and implementation of RR, describe its performance on a variety of workloads, and identify constraints on hardware and operating system design required to support our approach. * Majority of work carried out while supported by Mozilla Research.

show abstract

“…They designed blueprint-a set of abstractions, general principles, and low-level implementation details-for efficient deterministic replay in a modern hypervisor. They evaluated the proposed architecture in Xen, a full-featured hypervisor [23]. Trusted computing in cloud is also supported by layered attestation services.…”

Section: Recent Challenges : Trust and Virtualization In The Cloudmentioning

confidence: 99%

Convergence in trusted computing and virtualized systems: A new dimension towards trusted intelligent system

Vinh¹,

Bouzefrane²,

Banerjee³

2016

2016 International Conference on Performance Evaluation and Modeling in Wired and Wireless Networks (PEMWN)

View full text Add to dashboard Cite

The advantages of virtualization are tremendous in the current computing ecosystem. The virtualization technology is an abstract mechanism that enables a single hardware platform to run multiple environments. This technology is presented in many forms such as process, storage, and network virtualization. In this paper, we introduce the different techniques of virtualization such as those implemented by hypervisors or by isolated containers, in order to investigate the convergence between trusted computing and virtualized systems. By introducing Trusted Platform Module (TPM) in the virtualized environment, we describe the way VMs are attested to be used as a building block for any trusted intelligent system.

show abstract

Abstractions for Practical Virtual Machine Replay

Cited by 10 publications

References 20 publications

McFly: Time-Travel Debugging for the Web

McFly: Time-Travel Debugging for the Web

Engineering Record And Replay For Deployability: Extended Technical Report

Convergence in trusted computing and virtualized systems: A new dimension towards trusted intelligent system

Contact Info

Product

Resources

About