A Year in Papers

We’ve reached the end of term again, and I’m taking a break from writing up papers over the holidays – a chance to replenish my backlog and start planning for 2016 too! I want to see what I can do to improve the readability of the site as well. The Morning Paper will resume on the 4th January.

In a moment I’ll share with you the top 10 most read and most tweeted papers, plus some of my own picks. But first a quick look back over the year. Through the course of 2015 I’ve posted 206 paper write-ups on The Morning Paper plus a few original pieces and other miscellaneous posts. That means I’m now at over 300 paper reviews in total since #themorningpaper began. It’s amazing how a little every day adds up over time!

I’d like to say a huge thank-you to everyone who’s been following along, I love all the interaction that the papers lead to. And if you’re not yet subscribed to The Morning Paper and you’re looking for a New Years Resolution, signing up to the mailing list will get you approximately another 200 opportunities through 2016 to be exposed to something new and interesting, assuming I keep publishing at the same rate all though next year!

For those of you who are regular or semi-regular readers of The Morning Paper, I’d like to ask for your help in a couple of areas please:

  1. I’m often asked about the benefits of reading papers. I’d love to collect your stories – has a particular paper or idea within a paper influenced or inspired your work? It could be directly, but the indirect influences are also interesting. What do you personally get out of reading papers or the paper summaries on #themorningpaper? Do you share the summaries with others? (I’ve heard of industrial teams that forward the paper write-ups among all of their team members, and even researchers sharing paper summaries within their departments!) Please leave feedback either as comments on this blog post (I have to moderate comments because of spam, so it may take a while for them to appear), or via twitter, email, etc..
  2. There is so much wonderful work being done in research that deserves a wider audience. If you know of someone that might enjoy reading The Morning Paper, please help spread the word. Reading the papers and writing the posts is a lot of work (I calculated approximately how many hours I’ve spent on it this year, but the number is too scary to actually commit to type!), and it would be great to get the message out to as many people as possible.

Oh, and I also love hearing your suggestions and recommendations for papers you’d like to see covered too.

Before we dive into the top tens, a reminder that a good way to go back through the archives is to browse by month using the drop-down menu on the RHS of the blog (here’s January’s paper list for example).

My Top Picks of 2015

Picking out only 10 papers is really hard! Almost every week I find at least one paper that I’m really excited to share with you all, and I don’t write up any papers that I don’t like. Here are 10 I really enjoyed, and there are plenty more I’m sad not to be able to squeeze onto the list. I’m sure you’ll have your own favourites too.

Ordered by the date on which they first appeared on #themorningpaper…

  1. Viewstamped Replication Revisited
  2. Coordination Avoidance in Database Systems
  3. Muskeeter: all for one, one for all in data processing systems
  4. Scalability! But at what COST?
  5. The Dataflow Model: A Practical Approach to Balancing Correctness, Latency, and Cost in Massive-Scale, Unbounded, Out-of-Order Data Processing
  6. Feral Concurrency Control: An Empirical Investigation of Modern Application Integrity
  7. Mining your Ps and Qs: Detection of Widespread Weak Keys in Network Devices
  8. Holistic Configuration Management at Facebook
  9. IronFleet: Proving Practical Distributed Systems Correct
  10. From APIs to Languages: Generalising Method Names

How could I possibly miss out Unikernels, and Jitsu, and Molly, and Formal Methods at Amazon, and, The Scalable Commutativity Rule, and…, though!?

And a few original pieces which I was quite pleased with how they turned out:

The Top 10 Most Read Posts of 2015

These are the posts that had the most reads on the blog over the course of the year:

  1. Architecture of a Database System
  2. The O-Ring Theory of DevOps
  3. Twitter Heron: Stream Processing at Scale
  4. On the Duality of Operating System Structures
  5. Large-scale Cluster Management at Google
  6. Out of the Fire Swamp: Part I, see also Part II, Part III
  7. Scalability! But at what COST?
  8. Feral Concurrency Control: An Empirical Investigation of Modern Application Integrity
  9. From the Aether to the Ethernet: Attacking the Internet using Broadcast Digital Television
  10. Unikernels: Library Operating Systems for the Cloud

and just missing out on the top ten, at #11 we have Applying the Universal Scalability Law to Organisations

Top 10 on Twitter in 2015

How widely read a post is depends very much on the happenstance of someone posting a link to Hacker News, which noticeably bumps up the ratings (so feel free to post your favourite paper write-ups onto HN!). We can mitigate this factor in the top 10, by instead considering twitter activity relating to my paper announcement tweets

These are the posts that got the most retweets over the year. Remember, this may be influenced by how provocative my advertising tweet was!

  1. Feral Concurrency Control: An Empirical Investigation of Modern Application Integrity
  2. Twitter Heron: Stream Processing at Scale
  3. FaRM: Fast Remote Memory
  4. Optimizing Hash-Array Mapped Tries for Fast and Lean Immutable JVM Collections
  5. Scalability! But at what COST?
  6. Zab: High-performance broadcast for primary-backup systems
  7. Muskeeter: all for one, one for all in data processing systems
  8. Asynchronous Complex Analytics in a Distributed Dataflow Architecture
  9. IronFleet: Proving Practical Distributed Systems Correct
  10. Knowledge and Common Knowldedge in a Distributed Environment

If we order by likes instead, the following also make the top 10: