WebAssembly as an Enabler for Next Generation Serverless Computing

被引:9
|
作者
Kjorveziroski, Vojdan [1 ]
Filiposka, Sonja [1 ]
机构
[1] Ss Cyril & Methodius Univ, Fac Comp Sci & Engn, Rudzer Boshkovikj 16, Skopje 1000, North Macedonia
关键词
Serverless computing; WebAssembly; Function as a service; Internet of things; Performance evaluation; Benchmarks; WEB;
D O I
10.1007/s10723-023-09669-8
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
WebAssembly is a new binary instruction format and runtime environment capable of executing both client side and server side workloads. With its numerous advantages, including drastically reduced cold start times, efficiency, easy portability, and compatibility with the most popular programming languages today, it has the potential to revolutionize serverless computing. We evaluate the impact of WebAssembly in terms of serverless computing, building on top of existing research related to WebAssembmly in cloud and edge environments. To this end, we introduce a novel benchmarking suite comprised of 13 different functions, compatible with WebAssembly, and focusing on both microbenchmarking and real-world workloads. We also discuss possibilities of integrating WebAssembly runtimes with the application programming interfaces and command line interfaces of popular container runtimes, representing an initial step towards potential reuse of existing orchestration engines in the future, thus solving the open issue of WebAssembly workload scheduling. We evaluate the performance of such an integration by comparing the cold start delays and total execution times of three WebAssembly runtimes: WasmEdge, Wasmer, and Wasmtime to the performance of the containerd container runtime, using distroless and distro-oriented container images. Results show that WebAssembly runtimes show better results in 10 out of 13 tests, with Wasmtime being the fastest WebAssembly runtime among those evaluated. Container runtimes still offer better compute performance for complex workloads requiring larger execution times, in cases where cold start times are negligible compared to the total execution time.
引用
收藏
页数:20
相关论文
共 50 条
  • [21] NEXT-GENERATION COMPUTING
    KOEMAN, H
    IEEE SPECTRUM, 1984, 21 (01) : 8 - 8
  • [22] NEXT-GENERATION COMPUTING
    AVIZIENIS, A
    IEEE SPECTRUM, 1984, 21 (02) : 10 - 10
  • [23] Next generation reservoir computing
    Gauthier, Daniel J.
    Bollt, Erik
    Griffith, Aaron
    Barbosa, Wendson A. S.
    NATURE COMMUNICATIONS, 2021, 12 (01)
  • [24] Next generation reservoir computing
    Daniel J. Gauthier
    Erik Bollt
    Aaron Griffith
    Wendson A. S. Barbosa
    Nature Communications, 12
  • [25] NEXT-GENERATION COMPUTING
    BATE, R
    IEEE SPECTRUM, 1984, 21 (03) : 6 - 6
  • [26] NEXT-GENERATION COMPUTING
    PRESTON, K
    IEEE SPECTRUM, 1984, 21 (05) : 8 - &
  • [27] The Serverless Trilemma Function Composition for Serverless Computing
    Baldini, Ioana
    Cheng, Perry
    Fink, Stephen J.
    Mitchell, Nick
    Muthusamy, Vinod
    Rabbah, Rodric
    Suter, Philippe
    Tardieu, Olivier
    PROCEEDINGS OF THE 2017 ACM SIGPLAN INTERNATIONAL SYMPOSIUM ON NEW IDEAS, NEW PARADIGMS, AND REFLECTIONS ON PROGRAMMING AND SOFTWARE (ONWARD!'17), 2017, : 89 - 103
  • [28] The Rise of Serverless Computing
    Castro, Paul
    Ishakian, Vatche
    Muthusamy, Vinod
    Slominski, Aleksander
    COMMUNICATIONS OF THE ACM, 2019, 62 (12) : 44 - 54
  • [29] The rise of serverless computing
    Castro, Paul
    Ishakian, Vatche
    Muthusamy, Vinod
    Slominski, Aleksander
    Communications of the ACM, 2019, 62 (12): : 44 - 54
  • [30] Survey on serverless computing
    Hassan B. Hassan
    Saman A. Barakat
    Qusay I. Sarhan
    Journal of Cloud Computing, 10