RPCValet: NI-driven tail-aware balancing of µs-scale RPCs Daglis et al., ASPLOS'19 Last week we learned about the [increased tail-latency sensitivity of microservices based applications with high RPC fan-outs. Seer uses estimates of queue depths to mitigate latency spikes on the order of 10-100ms, in conjunction with a cluster manager. Today’s paper choice, RPCValet, operates at … Continue reading RPCValet: NI-driven tail-aware balancing of µs-scale RPCs
Author: adriancolyer
Venture Partner with Accel in London, author of 'The Morning Paper'
Understanding real-world concurrency bugs in Go
Understanding real-world concurrency bugs in Go Tu, Liu et al., ASPLOS'19 The design of a programming (or data) model not only makes certain problems easier (or harder) to solve, but also makes certain classes of bugs easier (or harder) to create, detect, and subsequently fix. Today’s paper choice studies concurrency mechanisms in Go. Before we … Continue reading Understanding real-world concurrency bugs in Go
Seer: leveraging big data to navigate the complexity of performance debugging in cloud microservices
Seer: leveraging big data to navigate the complexity of performance debugging in cloud microservices Gan et al., ASPLOS'19 Last time around we looked at the DeathStarBench suite of microservices-based benchmark applications and learned that microservices systems can be especially latency sensitive, and that hotspots can propagate through a microservices architecture in interesting ways. Seer is … Continue reading Seer: leveraging big data to navigate the complexity of performance debugging in cloud microservices
An open-source benchmark suite for microservices and their hardware-software implications for cloud & edge systems
An open-source benchmark suite for microservices and their hardware-software implications for cloud & edge systems Gan et al., ASPLOS'19 Microservices are well known for producing ‘death star’ interaction diagrams like those shown below, where each point on the circumference represents an individual service, and the lines between them represent interactions. Systems built with lots of … Continue reading An open-source benchmark suite for microservices and their hardware-software implications for cloud & edge systems
Distributed consensus revised – Part III
Distributed consensus revised (part III) Howard, PhD thesis With all the ground work laid, the second half of the thesis progressively generalises the Paxos algorithm: weakening the quorum intersection requirements; reusing intersections to allow decisions to be reached with fewer participants; weakening the value selection rules; and sharing phases to take best advantage of the … Continue reading Distributed consensus revised – Part III
Distributed consensus revised – Part II
Distributed consensus revised (part II) Howard, PhD thesis In today’s post we’re going to be looking at chapter 3 of Dr Howard’s thesis, which is a tour ("systematisation of knowledge", SoK) of some of the major known revisions to the classic Paxos algorithm. Negative responses (NACKs) In classic Paxos acceptors only send replies to proposer … Continue reading Distributed consensus revised – Part II
Distributed consensus revised – Part I
Distributed consensus revised Howard, PhD thesis Welcome back to a new term of The Morning Paper! To kick things off, I’m going to start by taking a look at Dr Howard’s PhD thesis, ‘Distributed consensus revised’. This is obviously longer than a standard paper, so we’ll break things down over a few days. As the … Continue reading Distributed consensus revised – Part I
End of term
We've reached the end of term again on The Morning Paper, and I'll be taking a two week break. The Morning Paper will resume on Tuesday 7th May (since Monday 6th is a public holiday in the UK). My end of term tradition is to highlight a few of the papers from the term that … Continue reading End of term
Keeping master green at scale
Keeping master green at scale Ananthanarayanan et al., EuroSys'19 This paper provides a fascinating look at a key part of Uber’s software delivery machine. With a monorepo, and many thousands of engineers concurrently committing changes, keeping the build green, and keeping commit-to-live latencies low, is a major challenge. This paper introduces a change management system … Continue reading Keeping master green at scale
Teaching rigorous distributed systems with efficient model checking
Teaching rigorous distributed systems with efficient model checking Michael et al., EuroSys'19 On the surface you might think today’s paper selection an odd pick. It describes the labs environment, DSLabs, developed at the University of Washington to accompany a course in distributed systems. During the ten week course, students implement four different assignments: an exactly-once … Continue reading Teaching rigorous distributed systems with efficient model checking
the morning paper 