当前位置: X-MOL 学术ACM Trans. Comput. Syst. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Derecho
ACM Transactions on Computer Systems ( IF 1.5 ) Pub Date : 2019-04-04 , DOI: 10.1145/3302258
Sagar Jha 1 , Jonathan Behrens 2 , Theo Gkountouvas 1 , Matthew Milano 1 , Weijia Song 1 , Edward Tremel 1 , Robbert Van Renesse 1 , Sydney Zink 1 , Kenneth P. Birman 1
Affiliation  

Cloud computing services often replicate data and may require ways to coordinate distributed actions. Here we present Derecho, a library for such tasks. The API provides interfaces for structuring applications into patterns of subgroups and shards, supports state machine replication within them, and includes mechanisms that assist in restart after failures. Running over 100Gbps RDMA, Derecho can send millions of events per second in each subgroup or shard and throughput peaks at 16GB/s, substantially outperforming prior solutions. Configured to run purely on TCP, Derecho is still substantially faster than comparable widely used, highly-tuned, standard tools. The key insight is that on modern hardware (including non-RDMA networks), data-intensive protocols should be built from non-blocking data-flow components.

中文翻译:

德雷乔

云计算服务经常复制数据,并且可能需要协调分布式操作的方法。在这里,我们介绍 Derecho,一个用于此类任务的库。API 提供了将应用程序结构化为子组和分片模式的接口,支持其中的状态机复制,并包括有助于在故障后重新启动的机制。运行超过 100Gbps 的 RDMA,Derecho 可以在每个子组或分片中每秒发送数百万个事件,并且吞吐量峰值达到 16GB/s,大大优于以前的解决方案。配置为纯粹在 TCP 上运行,Derecho 仍然比同类广泛使用的、高度调整的标准工具快得多。关键的见解是,在现代硬件(包括非 RDMA 网络)上,数据密集型协议应该由非阻塞数据流组件构建。
更新日期:2019-04-04
down
wechat
bug