9 Comments

Summary:

Skype went down this morning, plunging millions of users into VoIP and IM darkness. Then Twitter was out briefly before coming back online. The incidents may be unrelated, but they highlight our growing reliance on web communications tools to keep in touch.

2867982995_be003473ce_z

Today is the kind of day that make a user question his or her willingness to rely on cloud communications tools. Skype went down this morning, plunging millions of users into phone and IM darkness. Then, Twitter was out briefly before coming back online. The incidents may be unrelated, but they highlight our growing reliance on web communication tools to keep in touch. As we move our exchanges to the cloud, we are increasingly vulnerable to these kinds of outages.

We here at GigaOM were affected when we couldn’t reach contacts via Skype. Our development team coordinates on Skype, and was forced to look to other channels. One alternative was Twitter, which has become an important tool with DM (direct messages). That worked for my colleague Stacey until Twitter went down. A Facebook outage today could have had us preparing for the end of the world.

This is obviously not the first time something like this has happened. We wrote about a big Skype outage in 2007, wondering if it raised concerns about the reliability of P2P communications networks. But a few years later, we’re not just relying on Skype. So many of our ways of communicating are heavily dependent on cloud applications. When one goes down, it can strain the other channels we turn to pick up the slack. Last year’s Gmail outage also slowed down Twitter as people turned to that platform to vent and communicate. Today’s Skype outage was trumpeted by many on Twitter, which was also the tool Skype used to update its users. Imagine if Google bought Twitter or Facebook bought Skype. We’d have even fewer alternatives to route our traffic.

The lesson here isn’t that we should abandon online communications tools. Those are here to stay. We need to just keep our options open and be prepared to roll with the punches. Outages are going to be with us for many years to come, especially as these services scale up. The key thing is to have multiple channels ready if and when the next Skypefail or Twitter outage occurs. Think of it like an emergency back-up plan, much like families have rally points to meet up in case of disaster. For example: Meet on Twitter if Gmail fails. It may sound extreme, but with more and more companies and start-ups relying on these tools, it doesn’t hurt to prepare for the worst.

Related content from GigaOM Pro (sub req’d):

  1. Skype’s failure today doesn’t make me question “cloud communications tools” in general. I realize everybody has their own definition, but email runs “in the cloud” and it’s working fine. The Public Switched Telephone Network is perhaps the oldest “cloud” service and it’s doing fine today; the mobile service providers all seem to be holding up fine.

    Networks and applications that are carefully and rigorously designed, implemented, tested and maintained can be very reliable. But these days, there’s plenty of pressure to deploy and enhance and expand, and that often means that the newer services are less reliable.

    It sounds like Skype’s problems were perhaps due to a bug they introduced with a recent software update. In the old paradigm, that update would have probably gone through months of testing, before being deployed selectively to get more “real world” time before going mainstream, always with a quick way to revert if a problem were discovered.

    What perhaps should be re-thought is a choice to “rely” on rapidly-evolving services where the providers may tilt the balance away from stability and reliability in favor of feature innovation and growth.

    As users of these services, we make those choices and live with the consequences.

    I have great respect for many of the legacy “utility” providers who manage to deliver terrific up-time without a lot of fanfare. That would include not just communications, but power and water and sewer. Talk about keeping things flowing…

    Share
  2. THINK before you make the jump to “the cloud” What are you risking by rushing to outsource your critical infrastructure?

    What if the cloud provider doesn’t or your ISP is down? See Skype/Twitter issues above What are your employees doing while your connection to “the cloud” isn’t available?

    What if the cloud provider updates their systems and inadvertently causes losses due to employee retraining needs and/or corrupts your critical data?

    What if the police show up at the cloud provider with a cease and desist order for whatever reason? Sure, maybe you are smart and have a copy of your data but what good will it do you if you can’t run the program?

    Plan for failure. It will happen.

    Share
  3. As of 8:26am December 23rd on the West Coast of the United States, it’s up but a little spotty.

    I guess we’ll have to use email.

    Share
  4. [...] basic, daily operations such as remote teleconferencing. As those companies scrambled yesterday, GigaOm highlighted the trend with its piece, “Is Your Internet Disaster Plan in Place?” For alternatives to Skype [...]

    Share
  5. It used to be called Disaster Recovery Planning but now more commonly referred to as Business Continuity Services and a practice widely used in the enterprise where the cost of downtime exceeds the cost to plan for it. As the use of cloud services grows and becomes more mission critical, the need for cloud-based BCS options will grow as well.

    Share
  6. [...] back to life today after its first major outage in three years, it’s helpful to look at how critical the communications tool has become for users and companies, who increasingly rely on it for phone, IM and video chat [...]

    Share
  7. I have to agree that the real problem is the current state of the software development industry. These days, reliability is not a priority, especially among start-ups. Basically, people are getting used to junk software, much as they are willing to tolerate the downsides of junk food. Having your site crash is almost a badge of honor the clowns running start-ups and VC firms these days.

    Facebook and Twitter are making a fortune on junk software. When their services go down or screw up, they post cheap mea culpa’s, and their mindless users kow tow and thank them for their “honesty” or “integrity”.

    There was a time when updates were thoroughly tested before they ever made it to a production server. Now, there’s this bizarre cult of pushing to production several times a day, as if that’s some kind of demonstration of one’s machismo.

    It’s not going to get better, especially as long as the industry is run by children and psycho VC’s.

    Share
  8. An important part of the disaster plan is to understand how your critical infrastructure (Skype, in this case) is implemented. Before the outage that was a pretty opaque thing for a lot of people, and will probably continue to be with Skype. But you can bet that for people with much more critical communcation needs they do an immense amount of due diligence on their communication providers if they deem it mission critical.

    Share
  9. [...] If anything, 2010 was the year when our reliance on web services went through the roof. Whether it was social networking on Facebook, calling our friends on Skype or remixing content on Tumblr, we spent an inordinate amount of time on the Internet. And just as our usage went up, so did our expectations of uptime and reliability of these services. [...]

    Share

Comments have been disabled for this post