• What
    is this?

    You've landed on the AMD Portal on AnandTech. This section is sponsored by AMD. It features a collection of all of our independent AMD content, as well as Tweets & News from AMD directly. AMD will also be running a couple of huge giveaways here so check back for those.

    PRESENTED BY

The Facebook Server

In the basement of the Palo Alto, California headquarters, three Facebook engineers built Facebook's custom-designed servers, power supplies, server racks, and battery backup systems. The Facebook server had to be much cheaper than the average server, as well as more power efficient.

The first change they made was the chassis height, going for a 1.5U high design as a compromise between density and making the server easier to cool. 1.5U allows them to use taller heatsinks, larger (60mm) lower-RPM fans than the screaming 40mm energy hoggers used in a 1U chassis. The result is that the fans consume only 2% to 4% of the total power, which is pretty amazing as we have seen 1U fans that can consume up to one third of the total system power. It seems that air-cooling in the Open Compute 1.5U server is as efficient as the best 3U servers.

At the same time, Facebook Engineering kept the chassis very simple, without any plastic. It makes the airflow through the server smoother and reduces weight. The bottom plate of one server serves as the top plate for the server beneath it.

Facebook has designed an AMD and an Intel motherboard, both manufactured by Quanta. Much attention was paid to the efficiency of the voltage regulators (94% efficiency). The other trick was again to remove anything that was not absolutely necessary. These motherboards have no BMC, very few USB (2) and NIC ports (2), one expansion slot, and are headless (no videochip).

The only thing that an administrator can do remotely is "reboot over LAN". The idea is that if that does not help, the problem is in 99% of cases severe enough that you have to send an administrator to the server anyway.

The AMD servers are mostly used as Memcached servers, as the four channels of AMD Magny-cours Opterons 6100 are capable of using 12 DIMMs per CPU, or 24 DIMMs in total. That works out to 384GB of caching memory.

In contrast the Facebook Open Compute Xeon servers only have six DIMM slots as they are used for processing intensive tasks such as the PHP "assembling" data servers.

Cloud Computing = x86 and Open Source The Power Supply
POST A COMMENT

62 Comments

View All Comments

  • fpsvash - Thursday, November 03, 2011 - link

    In the middle of the paragraph below the image caption, the sentence reads "...and offers better slightly better performance..."

    Other than that, nice post!
    Reply
  • InternetGeek - Thursday, November 03, 2011 - link

    It's interesting that no many players have taken a look at Open Compute. Reply
  • alent1234 - Thursday, November 03, 2011 - link

    it's a solution for a specific workload. there are still a lot of workloads that require the traditional model of big database servers

    unlike your bank, facebook's noSQL is not ACID
    Reply
  • FunBunny2 - Saturday, November 05, 2011 - link

    Well, yes a voice of reason. OTOH, the Facebook et al folks are convinced that their back to the COBOL era is the future. As if a toy application, albeit pervasive, is "innovation". Reply
  • Sivar - Saturday, November 05, 2011 - link

    It's a little difficult to look at a comment about Facebook being a toy application and take it seriously. Yes, Facebook is not directly processing bank transactions on a Tandem, but their site is used to conduct business -- and is even the basis for many businesses, all over the world.

    Zynga, the company that makes a few annoying games for Facebook, is worth $15 -- more than Electronic Arts.

    Nearly every major online publisher, including Anandtech, uses their API for content distribution and often as the entire forum system for discussion of publications.

    The founder is the youngest billionaire in history.

    Calling theirs a toy application sounds like a Blockbuster customer calling Redbox a toy. It's denial of an obviously successful, large, powerful, innovative company because they don't do things "the old way."

    I suspect what matters more is that the business is executing flawlessly, the actual problems with data loss or other non-ACID compliant traditional issues are minimal, and that they are making enough money that Google and Microsoft are feel seriously threatened.

    One last thing -- if you really look into what ACID compliance means (and I know you did not specifically mention the acronymn, but replied to someone that did) none of the current major DBMS's are truly ACID compliant. It's too slow. Not Oracle. Not MSSQL. Not Greenplum. Not Teradata. None of them. They may be closer than NoSQL or the like, but then it's all about the right tool for the job, right?
    Reply
  • Ceencee - Wednesday, November 09, 2011 - link

    This is true but ACID can be over-rated for many workloads. How many pieces of data HAVE to be consistent across the entire cluster to be valid? What about NoSQL with configurable consistency like Cassandra?

    NoSQL databases provide the holy grail of system growth which is horizontal scaling and this is no small thing for anyone who has worked with a very large RDBMS like ORACLE and implemented RAC to find it doesn't scale all that linearly for most workloads.
    Reply
  • ac2 - Thursday, November 03, 2011 - link

    Wouldn't the presence of the graphics on the HP server account for the 32W idle load savings? Reply
  • JohanAnandtech - Thursday, November 03, 2011 - link

    It is an ATI ES 1000, that is a server/thin client chip. That chip is only 2D. I can not find the power specs, but considering that the chip does not even need a heatsink, I think this chip consumes maybe 1W in idle. Reply
  • mczak - Thursday, November 03, 2011 - link

    ES 1000 is almost the same as radeon 7000/ve (no that's not HD 7000...) (some time in the past you could even force 3d in linux with the open-source driver though it usually did not work). The chip also has dedicated ram chip (though only 16bit wide memory interface) and I'm not sure how good the powersaving methods of it are (probably not downclocking but supporting clock gating) - not sure if it fits into 1W at idle (but certainly shouldn't be much more). Reply
  • JohanAnandtech - Thursday, November 03, 2011 - link

    I can not find any good tech resources on the chip, but I can imagine that AMD/ATI have shrunk the chip since it's appearance in 2005. If not, and the chip does consume quite a bit, it is a bit disappointing that server vendors still use it as the videochip is used very rarely. You don't need a videochip for RDP for example. Reply

Log in

Don't have an account? Sign up now