Interesting technical facts about Facebook

Most of us use Facebook as social network application. Its famous, popular so it requires devotion and accuracy to manage all this. Following are some interesting facts

• #6 site on the internet
• 500 total employees
• 200 in engineering
• 25 in Infrastructure Engineering
• One of the largest MySQL installations in the world
• Big user and contributor to Memcached
• More than 10k servers in production
• 6,000 logical databases in production
  

Photo Storage and Management at Facebook:

• Photo facts:
• 6.5B photos in total
• 4 to 5 sizes of each picture is materialized (30B files)
• 475k images/second
• Mostly served via CDN (Akamai & Limelight)
• 200k profile photos/second
• 100m uploads/week
• Stored on netapp filers
• First level caching via CDN (Akamai & Limelight)
• 99.8% hit rate for profiles
• 92% hit rate for remainder
• Second level caching for profile pictures only via Cachr (non-profile goes directly against file handle cache)
• Based upon a modified version of evhttp using memcached as a “backing” store
• Since cachr is independent from memcachd, cachr failure doesn’t lose state
• 1 TB of cache over 40 servers
• Delivers microsecond response
• Redundancy so no loss of cache contents on server failure
  

Photo Servers

• Non-profile requests go directly against the photo-servers
• Only profile requests that miss the cachr cache.
• File Handle Cache (FHC)
• Based upon lighttpd and uses memcached as backing store
• Reduces metadata workload on NetApp servers
• Issue: filename to inode lookup is a serious scaling issue: 1) drives many I/Os or 2) wastes too much memory with a very large metadata caceh
• They have extended the Linux kernel to allow NFS file opens via inode number rather than filename to avoid the NetApp scaling issue.
• The inode numbers are stored in the FHC
• This technique offloads the NetApp servers dramatically.
• Note that files are write only. Mods write a new file and delete the old ones so the handles will fail and a new metadata lookup will be driven.
  

Issues with this architecture:

• Netapp storage overwhelmed by metadata (3 disk I/Os to read a single photo).
• The original design required 15 I/Os for a single picture (due to deeper directory hierarchy I’m guessing)
• Tracking last access time, last modified etc. has no value to Facebook. They really only need a blob store but they are using a filesystem at additional expense
• Heavy reliance on CDNs and caches such that netapp is basically almost pure backup
• 92% of non-profile and 99.8% of profile pictures are stored in CDN
• Many of the rest are almost all stored in caching layers
  

Solution: Haystacks

• Haystacks are a user level abstraction where lots of data is stored in a single file
• Store an independent index vastly more efficient than the file store
• 1M of metadata/1G of data
• Order of magnitude better on average than standard NetApp metadata
• 1 disk seek for all reads with any workload
• Most likely store in XFS
• Expect each haystack to be about 10G (with an index)
• Speaker equates a Haystack to be a lot like a LUN and could be implemented on a LUN. The actual implementation is via NFS onto NetApp as photos were previously stored
  

Net of what’s happening:

• Haystack always hits on the metadata
• Plan to replace NetApp
• Haystack is a win over NetApp but we’ll likely run over XFS (originally done by Silicon Grapics)
• Want more control of the cache behavior

Each Haystack Format:
Version number, Magic number, Length, Data, Checksum, Index format, Version, Photo key, Photo size, Start, Length.

• Not planning to delete photos at all since delete rate is VERY low so it the resource that would be recovered are not worth the work to recover them in the Facebook usage. Deletion just removes the entry from the index which makes the data unavailable but they don’t bother to actually remove it from the Haystack bulk storage system.

Q: Why not store the index in a RDBMS? Feels that it’ll drive too many I/Os and have the problems they are trying to avoid (I’m not completely convinced but do understand that simplicity and being in control has value).

• They still plan to use the CDN but they are hoping to reduce their dependence on CDN. They are considering becoming their own CDN (Facebook is absolutely large enough to be able to do this cost effectively today).

• They are considering using to SSDs in the future.

• Not interested in hosting with Google or Amazon. Compute is already close to the data and they are working to get both closer to users but don’t see a need/use for GAE or AWS at the Facebook scale.

• The Facebook default is to use databases. Photos are the largest exception but most data is stored in DBs. Few actions use transactions and joins though.

• Almost all data is cached twice: once in memcached and then again in the DBs.



Random bits:

• Canada: 1 out of 3 Canadians use Facebook.
• What is the strategy in China? A:“not to do what Google did” :-)
• Looking at de-duping and other commonality exploiting systems for client to server communications and storage (great idea although not clearly a big win for photos).
• 90% Indians access internet via a mobile device. Facebook very focused on mobile and international.

Read More

10 Must Have Softwares For A Blogger

This Post is Written by Karthik Kastury from dailyApps. You might want to subscribe to his feed if you liked this post. Please share any other softwares that you feel should make it to this list.

A Blogger is a person who is always on the search of something interesting and useful that is useful to his readers. This job can get a bit tedious if not done properly and without the use some special softwares made to make blogger’s life easier. Here are some of the softwares that a blogger must have inorder to get the best out of his blogging.

• Firefox Browser

Yes Firefox is the browser that I would suggest to all the blogger’s out there because of the presence of various addons and extensions to it and yes it is also Standards Compliant unlike IE.

• Del.icio.us Account
  

  Del.icio.us is a social bookmarking tool that helps you share links in a easy and effective manner. But there is much more to it. You Can use del.icio.us to discover new content for your blog and also keep track of everything you’ve discovered in the past so that you can retrieve it later easily.

• OpenDNS
  

  Yes OpenDNS is a very useful tool that not only helps you browse faster but also make it a lot more easier. For example you can create shortcuts to your most visited sites, and easily access them using OpenDNS. for eg. I use da to access this blog, gr to access google reader etc..

• An Offline Blog Editor
  

  Yes I would recommend either Windows Live Writer or thescribefire addon for firefox. Offline Blog Editors make blogging much more easier and also allow to manage multiple blogs at the same time.

• A Feed Reader
  

  A Feed reader is the bloggers best friend. A Feed reader allows you to keep track of your favorite blogs. Recommended feed readers are Google Reader or Bloglines.

• StumbleUpon Toolbar
  

  This is one must have toolbar. StumbleUpon allows you to discover some great new sites to blog about. StumbleUpon is my biggest source of content for this blog.

• Digg Account
  

  Digg is the posterboy of Web2.0. Aimed at providing the best Tech News, you can also use Digg to discover great content for your blog before everyone else blogs about it. Hat Tip useDigg Spy to discover stories that are worth a mention on your blog.

• Google Calendar & Google Notebook
  

  A Calendar is a must have tool for any blogger out there who wants to make it big. You should have a schedule set so that you don’t waste time on unnecessary things and at the same prioritize the things you need to do. Also a notebook tool where you can jot down important notes is also very important so that you don’t forget the next Killer idea. I Would recommend Google Calendar and Notebook for this.

• Firebug + YSlow
  

  This applies to all those bloggers who want to be Dugg. When you get dugg you’ll have huge burst of traffic and to make sure your blog does not go down fighting this mob you need to optimize your blog for minimal load times and server load. For this use Firebug and YSlow.

and last but not the least you definately need a blog to be a blogger right? Just Kidding. Hope this post will help you make yourself a much better and a productive blogger.

Read More

Raina powers India into Super Eights

GROS ISLET (Saint Lucia): Suresh Raina became only the third man to make a Twenty20 International century as India beat South Africa by 14 runs at the Beausejour Stadium on Sunday.

Scorecard

Victory, which saw India into the second round Super Eights after their seven-wicket win over Afghanistan on Saturday, was built on Raina's 101 out of a total of 186 for five.

Only West Indies' Chris Gayle and New Zealand's Brendon McCullum had previously made centuries at this level.

Raina had an escape on five when he was caught off a Morne Morkel no-ball.

It was an expensive error by the fast bowler, with left-hander Raina striking five sixes and nine fours in a blistering assault as he reached his century off 59 balls.

Together with Yuvraj Singh (37) he shared a stand of 88 in 10 overs after the pair came together with India 32 for two in the sixth over after they'd lost the toss.

South Africa captain Graeme Smith had promised his side would be flexible when it came to their openers but few foresaw the skipper dropping himself down to No 3.

Jacques Kallis, who opened in Smith's place, was nearly out for a second ball nought when a diving Yusuf Pathan at backward point just failed to hold on to a catch off the bowling of Harbhajan Singh.

Kallis, after the early loss of Loots Bosman, went on to complete a 45-ball fifty with three fours.

The experienced batsman then tried to keep South Africa on course by slog-sweeping Harbhajan for two successive sixes in the 15th over.

But a stand of 97 in 11 overs was ended when Smith was run out by Praveen Kumar's throw to wicketkeeper and India captain MS Dhoni.

Then 118 for two became 128 for three when Kallis fell on 73 after driving leg-spinner Piyush Chawla straight to Ravindra Jadeja in the deep.

South Africa though kept swinging and got the target down to 24 off the last over, bowled by Ashish Nehra and the left-arm quick conceded just nine.

India accelerated in the ninth over when Yuvraj swept spinner Roelof van der Merwe for the first six of the match and, two balls later, Raina lofted all-rounder Kallis high over wide long-on.

Both batsmen were severe on fast bowler Rory Kleinveldt, whose four overs cost 48 runs.

India took 25 runs off Kleinveldt's last over, the 18th, with Yusuf Pathan hitting the first ball for six and then taking a single.

Raina then struck three fours in a row, over long-on, through extra-cover and straight down the ground, before the batsman completed the over with a majestic six, struck off one knee and high over long-off.

Come the last over, Raina was on 95 and Dhoni was facing

But when he got on strike, for the third ball, Raina wasted little time in completing his hundred with a six over midwicket off Albie Morkel before getting out next ball.

This was only the 23-year-old Raina's 13th Twenty20 International and his score easily eclipsed his previous best of 61 not out against New Zealand in Christchurch last year.

Dhoni, whose side beat Afghanistan on Saturday, ended the innings with a six as India scored a mammoth 75 runs off the last five overs.

South Africa complete their group programme against Afghanistan in Barbados on May 5.

Read More

Top 10 private jets – Billionaires unlashed

When air travel is in question, the rich and famous fly in their own private jets. Equipped with luxurious interiors to suit their delicate taste, these private planes such as Cessna, Boeing, or Convair, make the trips their wealthy owners more comfortable and enjoyable. So, next time you fly in economy class I bet you’ll think about these lucky guys.
10. Donald Trump and his Boeing 727-23
Originally operated by American Airlines, this 1968 vintage jet was reconfigured to hold 23, with pale leather armchairs, gold plated seatbelt buckles, oil paintings, and Waterford crystal lamps. The “Trump” logo on the side of the aircraft is 30 feet long, 4 feet high, and made of 23 carat gold leaf.

9. Roman Abramovich and his Boeing 767-33A
This 767 may look ordinary on the outside, but its interior is reportedly outfitted with chestnut and decorated with gold. The aircraft can often be spotted at Luton Airport some 40 miles north of London, where Abramovich spends much of his time.

8. The Sultan of Brunei and his Boeing 747-430
The Sultan bought this 747 brand new for at least $100 million and had it fitted with a special interior and features such as washbasins of solid gold and Lalique crystal at an additional cost of some $120 million. The Sultan has several other aircraft, but this is his largest.

7. Jimmy Buffett and his Grumman HU-16 Albatross
This former military Grumman HU-16 Albatross amphibian aircraft owned by singer-songwriter Jimmy Buffett is named “The Hemisphere Dancer”. In 1996, it was shot at in Jamaica by local police who suspected it of carrying drugs.

6. Air Force One
The aircraft used to transport President George W. Bush on important state and domestic visits, Air Force One is a Boeing 747-200B that has been heavily modified with secure communications systems, electronic equipment, a self-contained baggage loader, front and aft air stairs, and the ability to refuel in-flight.

5. Mark Cuban and his Boeing 767-277
Billionaire Mark Cuban, who owns the Dallas Mavericks NBA basketball team, reportedly had custom-made seats installed on the aircraft that are large enough to accommodate the team’s tallest players.

4. Elvis Presley and his Convair 880
Named “Lisa Marie” after Elvis’ daughter, this Convair 880 was customized with 28 seats instead of the usual 110. The tail of the now preserved jet was painted with The King’s personal “TCB” logo, which stands for “Takin’ Care of Business”.

3. Bill Gates and his Bombardier BD-700 Global Express
Owned by Challenger Administration LLC on Bill Gates’ behalf, this jet can carry eight people at a cruising altitude of 51,000 feet for a distance of 6,500 nautical miles, a range that permits nonstop Tokyo-New York or Los Angeles-Moscow.

2. Wayne Huizenga and his de Havilland Canada DHC-6-320 Twin Otter
One of the most rugged and reliable utility transport aircraft ever built, the unpressurized Twin Otter can carry up to 19 people and take off from and land on rough strips as short as 100 yards. This aircraft appeared in a seaplane chase scene in the James Bond film “Casino Royale”.

1. Harrison Ford and his Cessna 525B CJ3 Citation Jet
Harrison Ford pilots his own CJ3, which can carry six people in comfort for some 1,900 nautical miles. Ford is considered so good a pilot that the FAA asked him to be the spokesman for the runway incursion awareness and prevention campaign that the agency started in 2001.

• If you happen to have one of these or maybe even better jet, please contact the administrator and you’ll be added shortly.

Read More