It’s not just you folks - if you haven’t seen the news on Twitter, Facebook is apparently having some service problems. Some people report that their pages have been loading after multiple attempts, but on the whole the largest social network in the world (300m+ users) is essentially down.
While most software applications only dream of having to deal with the level of scale Facebook does, it does point to a couple of key points in the transition to Software-as-a-Service (SaaS):
Multi-tenant Architecture
This is a fundamental premise for the business model to succeed for many SaaS vendors. Essentially, it is the use of the same software stack by multiple unassociated users. The challenges here are security and scale. Security is quite straight forward once a solid user registration and management system has been put in place. Scale is a bit trickier because it is dynamic.
Cloud Infrastructure Provider
This is the hardware (CPUs, memory, and storage) that the application runs on. Even though the architecture of the SaaS application could be developed to scale, the Cloud Infrastructure needs to be in place to support scale robustly (both the scaling up and the scaling down). There’s a reliance on the robustness of both the hardware and software architectures.
Facebook has their own datacenters. According to Datacenter Knowledge, as of April this year the Facebook datacenter stores more than 40 Billion photos and users upload nearly 40 million new photos per day! Now, that’s scale!
Multiple requests are in to FB for when they’re back up…hopefully it’s soon, or we might be forced into actually picking up the phone and talking to our mothers-in-law. 
UPDATE: Facebook is back up in San Francisco as of 18:00 PST. (Disaster averted!) But I see from Twitter that a lot of people are still having trouble with the platform.
UPDATE II: (18:15 PST) Facebook have disabled Chat - a clever way of freeing up some bandwidth for some? Many are still struggling to get back on the platform.
Posted under industry
