Amazon.com: Kindle DX: Amazon's 9.7" Wireless Reading Device (Latest Generation): Kindle Store

Kindle DX

Amazon.com: Kindle DX: Amazon's 9.7" Wireless Reading Device (Latest Generation): Kindle Store

Kindle DX

People who look exactly the same in all pictures (5 gifs) » Izismile.com - In fun we trust! Pictures, photos, videos, flash, games, celebs, hot stuff

People who look exactly the same in all pictures (5 gifs)

17 Apr 2009  / views: 348501

Do you remember the video with a girl who has the same pose on all her pictures? Well, I thought it was a one in a million case, but there are plenty of people who do this. O_o It seems that they think there’s only one side which looks good )
You can look through the photo album of your friends, maybe you'll notice the same thing

Chance of keeping your job if you’re WRONG more than 75% of the time

Chance of keeping your job if you’re WRONG more than 75% of the time

Graph by jpblue87, via our GraphJam builder.

Qcon

Shared by Madhu


Facebook architecture

Facebook Architecture Aditya Agarwal Director of Engineering 11/22/2008

The road to professional web development

http://www.slideshare.net/cheilmann/the-road-to-professional-web-development#

Facebook | Engineering @ Facebook's Notes

Scaling memcached at Facebook

by Paul Saab (notes) Saturday, December 13, 2008 at 1:13am

If you've read anything about scaling large websites, you've probably heard about memcached. memcached is a high-performance, distributed memory object caching system. Here at Facebook, we're likely the world's largest user of memcached. We use memcached to alleviate database load. memcached is already fast, but we need it to be faster and more efficient than most installations. We use more than 800 servers supplying over 28 terabytes of memory to our users. Over the past year as Facebook's popularity has skyrocketed, we've run into a number of scaling issues. This ever increasing demand has required us to make modifications to both our operating system and memcached to achieve the performance that provides the best possible experience for our users.

Because we have thousands and thousands of computers, each running a hundred or more Apache processes, we end up with hundreds of thousands of TCP connections open to our memcached processes. The connections themselves are not a big problem, but the way memcached allocates memory for each TCP connection is. memcached uses a per-connection buffer to read and write data out over the network. When you get into hundreds of thousands of connections, this adds up to gigabytes of memory-- memory that could be better used to store user data. To reclaim this memory for user data, we implemented a per-thread shared connection buffer pool for TCP and UDP sockets. This change enabled us to reclaim multiple gigabytes of memory per server.

Although we improved the memory efficiency with TCP, we moved to UDP for get operations to reduce network traffic and implement application-level flow control for multi-gets (gets of hundreds of keys in parallel). We discovered that under load on Linux, UDP performance was downright horrible. This is caused by considerable lock contention on the UDP socket lock when transmitting through a single socket from multiple threads. Fixing the kernel by breaking up the lock is not easy. Instead, we used separate UDP sockets for transmitting replies (with one of these reply sockets per thread). With this change, we were able to deploy UDP without compromising performance on the backend.

Another issue we saw in Linux is that under load, one core would get saturated, doing network soft interrupt handing, throttling network IO. In Linux, a network interrupt is delivered to one of the cores, consequently all receive soft interrupt network processing happens on that one core. Additionally, we saw an excessively high rate of interrupts for certain network cards. We solved both of these by introducing “opportunistic” polling of the network interfaces. In this model, we do a combination of interrupt driven and polling driven network IO. We poll the network interface anytime we enter the network driver (typically for transmitting a packet) and from the process scheduler’s idle loop. In addition, we also take interrupts (to keep latencies bounded) but we take far fewer network interrupts (typically by setting interrupt coalescing thresholds aggressively). Since we do network transmission on every core and since we poll for network IO from the scheduler’s idle loop, we distribute network processing evenly across all cores.

Finally, as we started deploying 8-core machines and in our testing, we discovered new bottlenecks. First, memcached's stat collection relied on a global lock. A nuisance with 4 cores, with 8 cores, the lock now accounted for 20-30% of CPU usage. We eliminated this bottleneck by moving stats collection per-thread and aggregating results on-demand. Second, we noticed that as we increased the number of threads transmitting UDP packets, performance decreased. We found significant contention on the lock that protects each network device’s transmit queue. Packets are enqueued for transmission and dequeued by the device driver. This queue is managed bv Linux’s “netdevice” layer that sits in-between IP and device drivers. Packets are added and removed from the queue one at a time, causing significant contention. One of our engineers changed the dequeue algorithm to batch dequeues for transmit, drop the queue lock, and then transmit the batched packets. This change amortizes the cost of the lock acquisition over many packets and reduces lock contention significantly, allowing us to scale memcached to 8 threads on an 8-core system.

Since we’ve made all these changes, we have been able to scale memcached to handle 200,000 UDP requests per second with an average latency of 173 microseconds. The total throughput achieved is 300,000 UDP requests/s, but the latency at that request rate is too high to be useful in our system. This is an amazing increase from 50,000 UDP requests/s using the stock version of Linux and memcached.

We’re hoping to get our changes integrated into the official memcached repository soon, but until that happens, we’ve decided to release all our changes to memcached on github.

Strategy: Facebook Tweaks to Handle 6 Time as Many Memcached Requests | High Scalability

Strategy: Facebook Tweaks to Handle 6 Time as Many Memcached Requests

Sat, 12/13/2008 - 21:19 — Todd Hoff

• Strategy: Facebook Tweaks to Handle 6 Time as Many Memcached Requests (810)

Our latest strategy is taken from a great post by Paul Saab of Facebook, detailing how with changes Facebook has made to memcached they have:

...been able to scale memcached to handle 200,000 UDP requests per second with an average latency of 173 microseconds. The total throughput achieved is 300,000 UDP requests/s, but the latency at that request rate is too high to be useful in our system. This is an amazing increase from 50,000 UDP requests/s using the stock version of Linux and memcached.

To scale Facebook has hundreds of thousands of TCP connections open to their memcached processes. First, this is still amazing. It's not so long ago you could have never done this. Optimizing connection use was always a priority because the OS simply couldn't handle large numbers of connections or large numbers of threads or large numbers of CPUs. To get to this point is a big accomplishment. Still, at that scale there are problems that are often solved.

Some of the problem Facebook faced and fixed:

Per connection consumption of resources. What works well at low number of inputs can totally kill a system as inputs grow. Memcached uses a per-connection buffer which adds up to a lot of memory that could be used to store data. Nothing wrong with this design choice, but Facebook made changes to use a per-thread shared connection buffer and reclaimed gigabytes of RAM on each server.

Kernel lock contention. Facebook discovered under load there was lock contention when transmitting through a single UDP socket from multiple threads. Sockets are data structures too and they are subject to the usual lock contention issues. Facebook got around this issue by maintaining separate reply sockets in different threads so they would not contend with the receive sockets. They found another bottleneck in Linux’s “netdevice” layer that sits in-between IP and device drivers. They changed the dequeue algorithm to batch dequeues so more work was done when they had the CPU.

Application lock contention. Nothing brings out lock issues like moving to more cores. Facebook found when they moved to 8 core machines a global lock protecting stats collection used 20-30% of CPU usage. In application that require little processing per request, as does memcached, this is not unexpected, but doing real work with your CPU is a better idea. So they collected stats on a per thread basis and then calculated a global view on demand.

Interrupt floods and starvation. With so much traffic directed at a single server the hardware can flood the CPU(s) with interrupts and keep the CPU from doing "real" work. To get around this problem Facebook implements some complicated strategies to load balance IO across all the cores. As I am less clever I might try more network cards with a TCP Offload engine.

When you read Paul's article keep in mind all the incredible number of man hours that went into profiling the system, not just their application, but the entire software hardware stack. Then add in the research, planning, and trying different solutions to see if anything changed for the better. It's a lot of work. Notice using a nifty new parallel language or moving to a cloud wouldn't have made a bit difference. It's complete mastery of their system that made the difference.

A summary of potential strategies:

Profile everything. Problems are always specific. The understanding of the problem must be specific. The fix must be specific.

Burn profiling into your regression tests. Detect when and where performance tanks as a regular part of your build.

Use resources in proportion to what grows slowest. This requires multiplexing, but at least your resource usage is more predictable and bounded.

Batch work. When you have the CPU do all the work you possibly can in the quantum or the whole system grinds to a halt in processing overhead.

Do work and maintain resources per task. Otherwise locking for shared resources takes more and more time when there's less and less time to do the work that needs to be done.

Change algorithms. Sometimes you simply need to do things differently. Tweaking will only get you so far.

You can find their changes on github, the hub that says "git."

Strategy: In Cloud Computing Systematically Drive Load to the CPU

Update: What are Amazon EC2 Compute Units?. Cloud providers charge for CPU time in voodoo units like "compute units" and "core hours." Geva Perry takes on the quest of figuring out what these mean in real life.

read more

Only His Only Grammarian Can Only Say Only What Only He Only Means.

Shared by Madhu


Placing the word "only" in 15 different places in the sentence results in over 20 different interpretations!

Only His Only Grammarian Can Only Say Only What Only He Only Means.

Peter G. Neumann, SEN Editor (1976-1994)
ACM SIGSOFT Software Engineering Notes 9, 1, Jan 1986, p. 6.
Copyright Peter G. Neumann, 1986,1997 [with slight revision in the placement of an italicized only -- see Note 2]

In response to the apparently rampant rage of the misplaced only,and perhaps also inspired by some of William Safire's columns in The NewYork Times Sunday Magazine, I offer the following annotated examples.(Italics indicate spoken stress, and parentheses indicate the presence ofsomething that would be implicit even when spoken.) Please bear with me onthis one. You may have to devote a little patience to each case. However,a conclusion you should be able to reach with no trouble at all is that theextent of ambiguity in the written English is awesome, but that wecan hide some of it in speaking the same words -- through suitable placementof emphasis (and with the help of a tolerance or ambivalence on the part ofthe listener). May you learn to put the only ``only'' only in thecorrect place. Note that the other places that you might put theitalicized only in the previous sentence (e.g., before the first,second, third, fourth, or tenth words) also result in different meanings.

1. Only(,) I said he thought secret users may write secret data.
2. Only I said he thought secret users may write secret data.
3. I only said he thought secret users may write secret data.
4. I said only (that) he thought secret users may write secret data.
5. I said only he thought secret users may write secret data.
6. I said he only thought secret users may write secret data.
7. I said he thought only (that) secret users may write secret data.
8. I said he thought only secret users may write secret data.
9. I said he thought only secret users may write secret data.
10. I said he thought secret users only may write secret data.
11. I said he thought secret users only may write secret data.
12. I said he thought secret users may only write secret data.
13. I said he thought secret users may write only secret data.
14. I said he thought secret users may write secret data only.
15. I said he thought secret users may write secret data only.

Note that ignoring the added stress and implicit parentheticals (as onewould in written English), various of these cases would appearindistinguishable from one another, reflecting inherent ambiguity in thewritten language; these cases are 1 and 2; 4 and 5; 8 and 9; 10 and 11; 14and 15. If you are still with me, here are some suggestions to help youinterpret at least one possible meaning in each of the above cases. Evenwith stress added, some of these cases are still ambiguous. And I have completely ignored the ambiguity between Secret-level usersand users of secrets, assuming the former for simplicity.

1. But I didn't say what you said I said. What I really said was ...
2. No one else said so. Only I did.
3. I didn't really mean it. I just said it.
4. I didn't say anything else.
5. I don't think anyone else thought so!
6. But he wasn't really sure.
7. He thought only that, and did not think anything else.
8. But not Unclassified or Top-Secret users.
9. But not Secret daemon processes or administrators.
10. Ambiguous! Depending on whether there is an unspoken pause or not after only, we have one of two cases:
The writer must be a user, and a Secret user at that, to write Secret data; OR
The only thing secret users may do is write Secret data (and nothing else -- not even breathe).
11. They are permitted to write, but perhaps they are unable to.
12. But they may not read such data.
13. But they many not write Top-Secret or Unclassified data.
14. They may not write secret procedures.
15. Ambiguous! Several of the above interpretations are possible, depending on substresses, pauses, and facial expressions.

========================================

NOTE 1, inspired by an Op-Ed piece,
Proposition 22 Author Didn't Take Grammar 101:
Improperly placed `only' opens Defense of Marriage Initiative to court challenge,
by Leo Kanowitz in the San Francisco Chronicle, 31 March 2000,
which cites the wording of Proposition 22 in the April 1999 statewideprimary election:

``Only marriage between a man and a woman will be valid or recognizedin California.''

Literally interpreted, that means that anything else between a man and awoman will not be valid or recognized, for example, legal contracts,spoken agreements, children, and a myriad of other things.It says nothing at all about the other relationships it was intended tomake invalid and unrecognized, such as same-sex partners, unions withanimals, etc. Very sloppy. The framers of Prop 22 should have readmy above discourse. PGN

========================================

NOTE 2, 19 April 2000: Many thanks to Julie Sussman, P.P.A., for noting theerroneous placement of the italicized ``only'' in the penultimate sentenceof the first paragraph, which has now been fixed above.

========================================