And by crippling, we really mean it – this multi-day fiasco was longest to disruption to BlackBerry service since it the device was introduced more than 10 years ago. The outages began Monday in Europe, Africa and the Middle East and have since spread through India and South America, reaching North America yesterday. The culprit? RIM’s centralized architecture that routes all BlackBerry traffic (web and email) through the company’s centralized infrastructure.
Jenna Wortham of the New York Times explained, “Because RIM sends its data through its own servers, any disruptions are felt by larger swaths of users than for other handset makers.”
Wortham’s referring to RIM and other companies in telecom, though the sentiment certainly applies to IAM. All services suffer outages from time to time – there’s no way around that. Modern architectures, however, can avert such catastrophic disturbances by decentralizing their architecture. That includes mobile, too.
That also includes Okta and the approach we take. In addition to being built on AWS (and spanning across a number of availability zones), our cloud-based service authenticates users, and then steps out of the way and lets browsers communicate directly to cloud apps. Older architectures, those not built natively in the cloud, rely on centralized proxies, similar to RIM.
When proxies work, life’s good. And when they don’t…well, multiple continents and millions of users feel the tremors.