I should certainly hope so!

Cyndy Aleo-Carreira509894 — Sun, 06 Jul 2008 14:19:09 -0700

I should certainly hope so! Working here is the second job I've gotten through working my LinkedIn contacts. ;)

About.com just released

Trisha — Wed, 02 Jul 2008 07:23:46 -0700

About.com just released their Top 10 List of the best employment websites:

http://jobsearch.about.com/od/joblistings/tp/jobbanks.htm

Linkedin made the list!

Facebook and LinkedIn talk scaling

Cyndy Aleo-Carreira — Tue, 01 Jul 2008 20:12:31 -0700

When one Web 2.0 application after another has trouble with scaling, (to the point where outages become as famous, if not more so, than the service) companies should be looking to emulate those examples that have achieved scale. Both Facebook and LinkedIn revealed a peek behind the curtains in two separate interviews that describe their scaling successes.

Facebook VP of Technical Operations Jonathan Heiliger spoke with CNET's Dan Farber about Facebook's architecture, which currently supports 80 million users, increasing at a rate of about 250,000 users per day. Using PHP as the primary development language, Facebook has a completely different challenge than Google does with scaling. Every query has to deal with the social relationships between users and data points with 50,000 transactions per second across the Facebook application as well as the platform, which hosts over 100,000 developers.

LinkedIn is far more forthcoming with their architecture, having presented a virtual how-to in two sessions employees gave at JavaOne 2008. Oren Hurvitz summarizes the slide shows (available at SlideShare), which give an intimate look at the company's numbers. LinkedIn has 22 million members, with over 4 million unique visitors each month. The site gets 40 million page views and 2 million searches per day.

What's even more interesting about LinkedIn's information is the timeline the company presented of the evolution of its architecture. When the site first launched (and for the first two years), engineers used a single Web app with one core database and a social graph cached in a cloud architecture. Five years later, that original web app is nothing more than the front end that users see when they log in; the actual nuts and bolts of LinkedIn are performed by broken-out services, each of which has its own database.

LinkedIn even went so far as to reveal the current hardware supporting "The Cloud," which caches the entire network graph: 22 million nodes with 40 instances. A rebuild of The Cloud from disk takes a full eight hours, and LinkedIn's estimates include support for up to 1.1 billion users before they would run out of RAM with their current set-up (Sun servers with up to 2 TB of RAM).

The LinkedIn presentations ended with tips any company should keep in mind when planning an application that needs to scale:

Can’t use just one database. Use many databases, partitioned horizontally and vertically.
Because of partitioning, forget about referential integrity or cross-domain JOINs.
Forget about 100% data integrity.
At large scale, cost is a problem: hardware, databases, licenses, storage, power.
Once you’re large, spammers and data-scrapers come a-knocking.
Cache!
Use asynchronous flows.
Reporting and analytics are challenging; consider them up-front when designing the system.
Expect the system to fail.
Don’t underestimate your growth trajectory.

Some companies may want to take out a pencil and notebook and write those down.

More news, commentary, and predictions from The Industry Standard:

Prediction: LinkedIn files for IPO in 2009
Analysis: How to safely use Facebook and LinkedIn at work
News: Facebook confirms Andreessen as its newest board member

The Industry Standard - Facebook and LinkedIn talk scaling - Comments

I should certainly hope so!

About.com just released

Facebook and LinkedIn talk scaling