Foundational distributed systems papers (2021) | Hacker News Reader

Foundational distributed systems papers (2021)

muratbuffalo.blogspot.com

193 pointsbelter24 comments

Past comments: https://news.ycombinator.com/item?id=26292627

This is an amazing resource - thank you for sharing. I did some undergrad research before landing my current position and really need to get back to exploring things outside of what I'm currently working with.

If anyone has any other papers to recommend for a junior engineer, please don't hesitate to share!

I think starting with books will help provide context and "the big picture" before jumping into the foundational papers.

Lots of folks recommend mixu's book: http://book.mixu.net/distsys/single-page.html

Thank you! This looks awesome.

I'm surprised not to see Raft on the list. I'm pretty new to the world, but it seems like it's become the replicated state-machine of choice.

There it is in the post:

In Search of an Understandable Consensus Algorithm. Diego Ongaro, John Ousterhout, Usenix ATC, 2014.

Oh duh. Sorry, I could have sworn "Raft" was in the title.

All is well here

Love to see stuff like this, where the whole foundation of a subject is collected into one resource. Thank you!

PSA: Make sure you read Butler Lampson's Hints for Computer System Design from the above list.

It is a classic paper(not specific to Distributed Computing) of Advice/Heuristics/Best practices for HW/SW design. Easy to read but quite timeless and insightful.

Public Service Announcement. Reading research papers is so important for your growth and career, please put in a process to do it at least once in a month or two. (I will try to write a blog post about why it is important, and how to go about it, since I see there is a big need for this.)

Papers We Love is a great resource, https://paperswelove.org/, if you like to get involved in a community to dip your feet into reading papers first.

And if you like to read papers physically https://mitpress.mit.edu/9780262045308/ideas-that-created-th... is a book of important papers in the field, complete with historical context.

Do you have an address of your blog? I'm interested in reading your blog post about your process

mad44 is I believe the author of the blog that OP linked, so the blog is just the OP, http://muratbuffalo.blogspot.com/ .

I don't see any direct articles about process as Murat characterizes himself as a researcher first and thus doesn't necessarily have much to say about “how do you fit this into your 9-to-5,” but he does give some reading advice,

"Deep Reading", http://muratbuffalo.blogspot.com/2022/02/deep-reading.html?m...

"Read papers, not too much, mostly foundational ones", http://muratbuffalo.blogspot.com/2021/02/read-papers-not-too...

Papers We Love was one of the two excellent meetups in NYC (the other being the Linux Users Group). That is, it wasn't just vendor advertisement time.

Unfortunately I don't think it's come back since the pandemic.

Hope it does.

While it's not the group, there is a Github repo which is constantly updated: https://github.com/papers-we-love/papers-we-love.

Github Repo for reference: https://github.com/papers-we-love/papers-we-love

Totally agree! If you're in Portland, join us for Papers We Love in person!

https://www.meetup.com/papers-we-love-pdx

Next meetup is Weds 10 May. We've been reading data-oriented papers recently (CRDTs/AutoMerge, Delta Lake, C-Store, etc) but we'll likely do some of these papers soon. Hope you can join us.

agreed. the best papers are often better reading than the vast majority of technical books or websites. perhaps because of the pressure to be clear and concise. with books I find I often must wade through lots of fluff or "novel to novices" ideas. best papers have a great S/N ratio

I've recently challenged myself to read more papers related to the work I do as an engineer. It was a muscle I exercised as an undergrad but have since neglected in my working life. I had imagined these papers to be much, much harder to parse than they really were, the domain of proper computer scientists, but that's just silly. It's akin to how one should read a primary source over a summary -- the original text is not a mystical text beyond your comprehension; you too can engage with it and draw your own conclusions.

> the original text is not a mystical text beyond your comprehension; you too can engage with it and draw your own conclusions.

This! This is so important.

Here is what I wrote (oh, wow, 10 years ago) about the process I use for reading papers. I hope it helps people.

https://muratbuffalo.blogspot.com/2013/07/how-i-read-researc...

On the flip side, if you don't have the foundation it's entirely possible you won't get that much out of reading the paper. I enjoy reading papers on databases, but I also know that unless I start actively working on DB internals a lot of the lower level details of a paper will escape me. So essentially when I'm reading a paper, unless it's something I'm intimately familiar with, I stick to the high level details.

If you or someone else is looking for a forum experience focused on discussing papers, try out https://paperlist.io/ - the creator of it posted paperlist here a while back and it's funny how I'll just catch myself browsing papers during breaks.

FWIW, It's got low levels of engagement comment-wise, but the site is only a few months old and papers are posted often

disclaimer: I don't know anybody involved with paperlist and am not involved with any aspect of paperlist other than reading and posting papers. I just like it.

As a researcher, this sounded interesting so I had a look. The first paper that caught my eye (https://paperlist.io/post/247381561) has a bunch of "generic" (lacking substance; could be generated) comments that I find a bit uncanny. Three of them have the same style of being all lower-case (even the name of R, the programming language being discussed). Doesn't make me want to engage.

... and now I also noticed that the paper summaries are AI-generated! That's an anti-feature for me, at present.

Yeah, after reading a few dozen comments on some papers I completely agree... the lack of substance is so clear and consistent it's highly suspicious.

The concept is interesting enough, but uh. If they're filling it with bot comments to make it look less empty to passers-by, they're optimizing for catching people not paying attention, while apparently targeting people interested in deep attention (who will immediately turn away from this).

Reply on news.ycombinator.com