Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The database servers powering Let's Encrypt

letsencrypt.org

Post reply on HN
Print view View original

211–220 of 236 posts

Re: The database servers powering Let's Encrypt

#211

Post by alberth »

tyingq wrote: post #25

I was, long ago, an old-school Unix sysadmin. While I was technically aware of how powerful smallish servers have become, this article really crystallized that for me. 64 cores and 24 NVME drives in a 2U spot on a rack is just insane compared to what we used to have to do to get a beefy database server. And it's not some exotic thing, just a popular mainstream Dell SKU. If you price it out on Dell's site, you get a r…

And if you don't want to buy Dell, you can get a SuperMicro server that has about 2x the specs for half the price ($102k).

https://www.siliconmechanics.com/system/a+-ultra-superserver...

2U

128-Core

307TB NVME

4 TB RAM

-------

= $102k

Re: The database servers powering Let's Encrypt

#212

Post by JohnJamesRambo »

yakubin wrote: post #210
kornbattery wrote: post #196

Earlier quoted context omitted.

That doesen't seem like some fundamental restriction, but rather a compromise we're always willing to make. 30ms is not noticeable, so we don't try to lower it and sacrifice something else. Even the super slow ones like on kindle, are _choices_ that have been made in favor of something else. a second to turn a page on a book isn't unbearable

30ms latency is acceptable. But that's the best we ever had. High-end Android phones have 120+ ms latecy. That's easily noticeable and actually annoying (at least to me). My personal pet peeve is the latency of input when i'm starting new applications in KDE. E.g. I'll start a new terminal with Super+Enter, followed by a Super+Right Arrow in order to tile it to the right. But the latency is big enough, that often it'…

It’s so frustrating that our phones and computers can’t move at the speed our mind moves. I know it could if someone cared to try to make it a priority.

Re: The database servers powering Let's Encrypt

#215

Post by JulianMorrison »

lisper wrote: post #113

I don't understand why they are trying so hard to avoid sharding. It seems to me that this is a perfect example of an "embarassingly parallel" problem for which sharding would be borderline trivial. What am I missing?

It's also an example of a "mustn't fail or you break the internet" problem and a "lots of people with nation state resources have a reason to fuck with us" problem. They are prioritising simplicity as a means to security, and that makes sense to me.

Re: The database servers powering Let's Encrypt

#216

Post by ddorian43 »

tanelpoder wrote: post #140
PeterCorless wrote: post #101

Earlier quoted context omitted.

We've made Scylla as async, shared nothing as possible, and we've also started adding C++20 coroutines (to replace futures/promises). We'll be doing more of that in 2021. https://www.scylladb.com/2021/01/12/making-scylla-a-monstrou...

I plan to test Scylla (and Postgres with TimescaleDB) out soon after done with basic Linux tests :)

Just note that TimescaleDB is olap (columnar compression), scylladb is oltp with async-fsync and Postgresql is disk-fsync-on by default.

Re: The database servers powering Let's Encrypt

#217

Post by giancarlostoro »

jaas wrote: post #104
tyingq wrote: post #25

I was, long ago, an old-school Unix sysadmin. While I was technically aware of how powerful smallish servers have become, this article really crystallized that for me. 64 cores and 24 NVME drives in a 2U spot on a rack is just insane compared to what we used to have to do to get a beefy database server. And it's not some exotic thing, just a popular mainstream Dell SKU. If you price it out on Dell's site, you get a r…

Just so everyone here is aware re: pricing... If you buy through a VAR and/or Dell reps you don't pay the price on the website. What you actually pay is typically significantly lower. I don't think anyone actually buys servers like these by just ordering from the website. We (Let's Encrypt) certainly don't. These are expensive servers, crossing into six digits, but not $200k.

Well, now I have to ask an old boss if he called a rep, he didnt spend quite 100k+ but we did get a twenty grand single server from dell about half a decade back or more, and now I'm curious, I think it was about 20 grand, I don't think NVMe's were mainstream at the time, or anywhere damn near affordable. I suspect he did buy it directly off the site, as he has other hardware we'd have gotten (mainly workstations).

In a way you could say the ones not calling a rep are subsidizing your cost for the server.

Re: The database servers powering Let's Encrypt

#218

Post by haltingproblem »

midasuni wrote: post #200
paxys wrote: post #153

Earlier quoted context omitted.

Literally the only reason to display a retail price on the website is so sales reps can say "see we are giving you a massive XX% discount!"

I wonder if such “price on application” behaviours is partially what drives services like AWS I’m looking at buying some fortigate firewalls to do some natting, looks like 200Fs will be fine, but even the fortigate sales guys won’t give me a price, I have to go to resellers who also refuse to give prices, which adds friction. Cisco exactly the same. When looking at options, price is at the forefront of my mind, but s…

You can back into the price from used gear.

Ebay has them for 3.2k and CDW new for 3.6.

Re: The database servers powering Let's Encrypt

#219

Post by haltingproblem »

alberth wrote: post #211
tyingq wrote: post #25

I was, long ago, an old-school Unix sysadmin. While I was technically aware of how powerful smallish servers have become, this article really crystallized that for me. 64 cores and 24 NVME drives in a 2U spot on a rack is just insane compared to what we used to have to do to get a beefy database server. And it's not some exotic thing, just a popular mainstream Dell SKU. If you price it out on Dell's site, you get a r…

And if you don't want to buy Dell, you can get a SuperMicro server that has about 2x the specs for half the price ($102k). https://www.siliconmechanics.com/system/a+-ultra-superserver... 2U 128-Core 307TB NVME 4 TB RAM ------- = $102k

We profiled Supermicro vs. Dell vs. HP vs. IBM servers a decade ago for trading applications.

Supermicro was cheaper by 20-30% had a quantitatively better power profile which basically meant it paid for itself and just rocked on every dimension. And their sales reps were able to answer every question we had including sending us a excel model of their power usage. The dell folks never got back to us on power usage.

I dunno why folks insist on Dell crapola.

Re: The database servers powering Let's Encrypt

#220

Post by driverdan »

njharman wrote: post #190
lisper wrote: post #113

I don't understand why they are trying so hard to avoid sharding. It seems to me that this is a perfect example of an "embarassingly parallel" problem for which sharding would be borderline trivial. What am I missing?

Rack space ain’t free

The cost of 4U vs 2U is trivial compared to the cost of this hardware.
Post reply on HN