Monthly Archives: November 2014

Data center fabric, the next-generation Facebook data center network and “How we scaled Instagram”, by Mike Krieger

Sh*ts
Introducing data center fabric, the next-generation Facebook data center network

The more than 1.35 billion people who use Facebook on an ongoing basis rely on a seamless, “always on” site performance. On the back end, we have many advanced sub-systems and infrastructures in place that make such a real-time experience possible, and our scalable, high-performance network is one of them.

Facebook’s production network by itself is a large distributed system with specialized tiers and technologies for different tasks: edge, backbone, and data centers. In this post, we will focus on the latest developments in our data center networking and unveil the next-generation architecture that we have successfully deployed in our new Altoona facility: data center fabric.

https://code.facebook.com/posts/360346274145943/introducing-data-center-fabric-the-next-generation-facebook-data-center-network

“How we scaled Instagram”, by Mike Krieger (video)

Mike Krieger, co-founder of Instagram, talks about their tech setup, how they scaled from one server in LA, to AWS and ultimately Facebook’s own infrastructure.

https://opbeat.com/blog/posts/how-we-scaled-instagram-mike-krieger/

 

rackspot changed his name to easyhost

Sh*ts

rackspot.pt changed his name to easyhost.pt … this is what happened…

they are now using online.net infrastructure… my blog, hosted on rackspot.pt – now on easyhost (online.net) is getting lot of database errors…. looks like they are even using the db connections limits…. fuck!

time to move!..
i already own a dedibox @ online.net … time to move it all.

UPDATE!! OK THEY WERE UNDER A DDOS ATTACK!
THATS WHY CONNECTIONS FAILED.

Fabric, the easiest way to build the best apps, from Twitter

Utilities,

 

Crashlytics

Introducing the most powerful, yet lightest weight crash reporting solution.

 

Beta by Crashlytics

The Most Seamless Beta Distribution Experience for You and Your Testers.

We applied the Crashlytics expertise to beta distribution. The result: a streamlined solution for distributing apps that gives you a single, cross-platform toolset for iOS and Android, and a delightful, effortless onboarding for your testers.

 

Sign in with Twitter

The fast and safe way for people to sign in to your app or website.

 

Growth, simplified.

Finally, user onboarding that helps you grow your app.

Tweet Embeds

The easiest way to bring Twitter content into your app.

Answers by Crashlytics

No Analysis Required.

Finally, mobile app analytics you don’t need to analyze. Built into our existing, award-winning Crashlytics Kit.

 

Mopub

Maximize your mobile ad revenue.

 

Raid 1 and Raid 0 explained

Sh*ts, , ,

 

I’m afraid you dont even get something as simple as:
Raid 1 : faast
Raid 0 : slow

Raid 1 brings safety by allowing you to half of screwed drives
Raid 0 brings you the comfort to know that when a drive is crashed, you can go home. To cry.

But speed wise it’s a different story:
Raid1:
– You have twice the same data, so write speed should be, let’s say, “normal
– You have 2 disks to read from. With the proper disks, controllers and bla bla, you read faster.

Raid 0:
– You write faster
– You read faster

So, in some cases where you require those little boosts, and the data is not “i’m going to kill myself if i lose this” important, then why not, just stack a few extra drives nearby ;)

In jflory’s case, a failed drive, would mean both drives are not useable, then waiting for the replacement, and timing … Yes that’s a lot of downtime for gaining a few G you might be able to save up anyway.

Dont be that guy Jflory!


http://www.spigotmc.org/threads/raid-1-raid-0-soyoustart-myth-busting.11171/#post-127898

Who remembers ARJ?

Sh*ts,

ARJ has a 10-plus year track record of good speed, compression and the most flexibility of any archiver. In addition, it will run on most PCs from the IBM PC XT to Pentium IIIs and is compatible from version 1.0 to present.

The major features of ARJ 2.86 include:

  • Windows 95, 98, Millenium long filename support in ARJ
  • Windows style wildcard support
  • Ability to execute a command after self-extraction of files
  • Automatic handling of more than 99 volumes
  • Ability to process and archive up to 50,000 files at one time (registered support for up to 260,000 files at one time)
  • Ability to store up to 250 backups (chapters) within a single archive and restore any specific backup
  • Excellent file compression in terms of resulting archive size and speed
  • Unmatched flexibility in archiver usage
  • Support for Windows file date-time created and accessed
  • Data damage protection option storing data within the archive
  • Improved chapter archive and multiple volume support
  • Year 2000 compliance