Jeff Darcy

Instead of repeating myself, I'll just cite the blog post I wrote about this, and hope the rest of HN doesn't mind. http://pl.atyp.us/2013-10-leaning-out.html Things have changed only a bit since then. A couple of years' worth of still-increasing income have brought me even closer to permanent financial security. I'm still pretty unimpressed by where this industry is headed technically, socially, and financially. I still want to finish one more Big Thing for pay, maybe one more on my own, and then just dabble in whichever smaller projects catch my fancy. I know, based on the responses to that blog post, that I'm far from alone. I suspect that many people reading this will eventually end up in the same place, whether they believe it now or not.

There's also an excellent writeup by GlusterFS contributor Jeff Darcy: http://pl.atyp.us/wordpress/index.php/2012/10/rackspace-block-storage/

For the record, his suggestion to openly commit to not use patents was taken up recently by Twitter and widely applauded, although he copped an amazing amount of flak at the time. I found his blog though, so all is not lost: http://pl.atyp.us/wordpress/ In this case, hellbanning just demonstrates that HN is run by a private cabal. Whoever it is can run it in their own way I guess, but without any oversight they're likely to ban people for the wrong reason and with no recourse, and doing so doesn't contribute to things like voting quality.

[Disclaimer: I am a committer on Project Voldemort] Redis is very cool (I really want to play with it - congrats to VMWare on hiring its author), but it is neither a disk-oriented system [EDIT: fixed per comment, was disk-persistent] nor (most importantly, IMO) a fully distributed system. See a comment in this blog post for great visual guide to these systems from these two axes (disk-orientation and distribution): http://pl.atyp.us/wordpress/?p=2929

Jeff Darcy tears into this one: http://pl.atyp.us/wordpress/?p=2555 "Letting fear of conventional storage drive the creation of 'solutions' that are just as complex without the benefits of being as general, as well tested, or as well documented is a mistake. (Open but undocumented and untested can be worse than closed, BTW, if the cost of reverse-engineering and fixing the implementation is greater than the cost of licensing would have been.) Such attempts generally lose even when considered alone, and even more so when the effects of fragmentation and incompatibility are considered."

Not to beat a dead horse or keep a totally unproductive thread alive longer than it has to be, but, if this is you: http://pl.atyp.us/ (from your profile) you seem like an interesting person with a lot of experience and perspective to contribute. So with that said: why not be specific about the selective over-application you're talking about? I couldn't find it in the comment history. People probably won't agree with you but it would probably be helpful for them to at least know where you're coming from. It has to be better than just resigning yourself to the idea that the moderators are working in bad faith against you. For my part, if it helps you to know this: I'm one of several people who have been leaning on the mods for literally years to add something about negativity to the guidelines. They pushed back on suggestions for a long time out of many of the exact same concerns people are voicing on this thread. It's not a sudden decision they've made.

I enjoyed this thought piece on technical debt vs technical rsik: http://pl.atyp.us/2015-01-technical-risk.html

Interesting couple of related articles / rants by Jeff Darcy: http://pl.atyp.us/2013-08-local-filesystems-suck.html http://pl.atyp.us/2013-11-fixing-fsync.html

In principle, yes. I believe most modern filesystems use (kernel-level) write barriers to ensure filesystem consistency, and those do depend on having hardware barriers as well. The kernel is not set up to export this interface, however, and there doesn't appear to be any serious work being done to fix that. EDIT: http://pl.atyp.us/2013-11-fixing-fsync.html has more details.

Ah, the myth of the sufficiently redundant network. http://pl.atyp.us/wordpress/index.php/2010/10/when-partitions-attack/ http://kellabyte.com/2013/11/04/the-network-partitions-are-rare-fallac…

This is cool but I think it's important to remember that they're comparing algorithmic benchmarks but modern software has many different bounds besides for raw, algorithmic processing. For example, I think I/O and database interactions are the bottleneck for most web apps. I really liked the way this article broke it down: http://pl.atyp.us/wordpress/?p=2947

Some perf numbers: Bitcask vs. InnoDB (by Dave Smith): http://twitter.com/dizzyco/status/13014285189 Bitcask vs. Tokyo Cabinet (by Jeff Darcy): http://pl.atyp.us/wordpress/?p=2868

I did say that I'd written about it on my site. Here are a couple of links. http://pl.atyp.us/wordpress/?p=2240 http://pl.atyp.us/wordpress/?p=2258 Here's one from Randy Bias, which I had also alluded to. http://cloudscaling.com/blog/cloud-computing/virtual-disks Now, why don't you take Dennis to task for his lack of supporting information? He provides even less information about methodology, and his results are simply not grounded in reality at all. In my experience, application of a double standard strongly implies preference for outcome. Why don't you actualy state your position, and then back it up to the same degree you demand others do? I know this is Kibitzer News and all, but it would be nice to see something besides sniping from the sidelines for once.

FWIW, the C10K Problem page (http://www.kegel.com/c10k.html) compares various approaches to handling tens of thousands of simultaneous clients. It's from 2006, but is a good overview of well-understood techniques. Also, some observations on server design from Jeff Darcy (http://pl.atyp.us/content/tech/servers.html).