A few years ago, a multi-day power failure on Long Island left our offices in the dark – including the Microsoft Exchange Server in our server closet. After that experience, and a few others involving the local electrical grid, we moved our mail system into the cloud for improved reliability.

This week, a failure of some sort took down one of the best-known cloud services. As I write this on Friday morning (April 22), Amazon is still struggling to fix Amazon Web Services, which has big problems in its Northern Virginia data center. Some services came back online on Thursday, but others are still down.

As of the time of writing, Amazon has not provided any sort of explanation for the outage, or disclosed how many of its customers were affected. The best customer reports are on Twitter. But according to the AWS Service Health Dashboard, there are issues with the company’s Electric Compute Cloud, Relational Database Service, Elastic MapReduce, Cloud Formation and Elastic Beanstalk resources in its US-EAST-1 region. The problems also include high latency and error rates, instance launches, API accesses, database connectivity… the list goes on and on.

Are/were you affected by the AWS failure? Probably. We were.

SD Times doesn’t use AWS for any of our internal IT needs. We have some systems that use Google’s Web services, but have nothing stored in Amazon, or that use the AWS APIs directly.

On the other hand, for our technical conferences (like SPTechCon for SharePoint developers and admins), we like to create funny promotional videos using Xtranormal. Those videos (again, as I write this) are down, and have been for 21 hours. As the company tweeted on Thursday, “Hang tight, peeps. Those pesky ‘Amazon’ issues you might have heard about this morning affected us as well.”

The good news is that we know one thing for sure that Skynet – the global computer system in the Terminator movies – didn’t take down Amazon Web Services. One of the company’s support team, Luke@AWS, posted,

From the information I have and to answer your questions, SkyNet did not have anything to do with the service event at this time.

The bad news is that, with the exception of purely siloed enterprise IT systems, we’re all using cloud computing, all the time. Even if we’re not subscribing to those services ourselves, the chain of dependencies, thanks to APIs, Web services, RSS feeds, and external data providers, means that our IT is only as strong as the weakest link.

Z Trek Copyright (c) Alan Zeichick
3 replies
  1. Craig
    Craig says:

    But thats how skynets starts!! One point of access and then spreads. Then the military is going to put Skynet online to tackle the problems; Thats when it gets interesting… 🙂

  2. Craig
    Craig says:

    That’s how it starts!! One point of access and it spreads. Once it has infiltrated more systems the US Air Force will flip Skynet online to reclaim the lost systems; thats when it gets interesting!!

  3. KonaKoder
    KonaKoder says:

    Isn’t it more likely that the AWS crash was caused by heroic time-traveling members of the Resistance? Indeed, the AWS crash SAVED humanity?

Comments are closed.