TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Twitter, PayPal reveal database performance

119 点作者 werner超过 12 年前

11 条评论

dkhenry超过 12 年前
I always knew the company I work for does cool things with MySQL, but I never knew how far we really have pushed the envelope until I see how much more we push MySQL then the twitters and facebooks of the world. Doing some napkin math I would say our largest cluster of MySQL does about 320 Billion inserts a Day across 80 nodes and some of our nodes can peak at about 200K queries per second ( that would be 16 Million QPS if all the nodes were fully loaded ). Meaning at one of our installs we do the same volume as two facebooks !
评论 #4616801 未加载
评论 #4616496 未加载
ericcholis超过 12 年前
Not surprising, especially if you've ever used their interface or API. I'm reasonably certain that they don't have any indexes.<p>Slowest API ever....
评论 #4616357 未加载
评论 #4616832 未加载
评论 #4616450 未加载
评论 #4616458 未加载
robryan超过 12 年前
The speed of their transaction search actively discourages me using it. It is extremely slow even when searching by the transaction Id which you would assume at least that would have a decent index on it.
评论 #4616467 未加载
aidos超过 12 年前
<i>"get a timestamp error if one node attempts to read data that was written by another node in the future (due to data moving across datelines)"</i><p>Honest question: Which part of a replication/db system would be using time stamps that weren't universal?
评论 #4616395 未加载
评论 #4616293 未加载
andrewcooke超过 12 年前
curiously, 4629 tweets a second is 4 million a day[1], which is more than the number of rows per day (3 million).<p>this must be someone not thinking, right? i don't see how they can have less rows than tweets. even with de-duplication of repeated text they still need a key (and timestamp) per tweet.<p>oh! and it gets worse. the next line says "400 million new tweets a day". that one must be plain wrong (it's a rate 100x higher than the number per second), given that the other two are (comparatively) consistent (it would also mean an average of 3 tweets per day per active user (140 million), and i suspect they define active user to be anyone over 1 tweet per month...).<p><i>so the section "by the numbers" contains four items, but only two independent values. and appears to be inconsistent twice.</i><p>[1] almost exactly - presumably the per second value is derived from a rough figure for the daily value.
评论 #4618384 未加载
评论 #4617248 未加载
评论 #4616491 未加载
gogetter超过 12 年前
Twitter, Paypal don't really need to be that fast. Nor does Facebook.<p>But do you think major banks on Wall Street tracking billions of trades a day in near real-time would be satifisfied with these numbers?<p>It all depends on what your users demand and what they will put up with.
spudlyo超过 12 年前
If anyone is interested in watching Jeremy's presentation online it can be found at:<p><a href="http://medianetwork.oracle.com/video/player/1873920417001" rel="nofollow">http://medianetwork.oracle.com/video/player/1873920417001</a>
keyz超过 12 年前
What I would really like to see is a guesstimate of the volume of user behavior/analytics data that FB/Twitter generate. It would be interesting to see the average amount of data per user-interaction.
VMG超过 12 年前
That's surprising to me. I didn't expect financial transactions to create that much data. Also wouldn't most of the data be archived and only rarely be accessed?
评论 #4617814 未加载
评论 #4616519 未加载
评论 #4616611 未加载
trustfundbaby超过 12 年前
It'd be great to read about people's experiences using postgres on a comparable scale to that discussed in this article ... does that even happen?
评论 #4617082 未加载
hendler超过 12 年前
Does the "commercial instance of MySQL" refer to Percona, something else, or is it simply the MySQL commercial license?
评论 #4619062 未加载