Chris Kranky

Recent Posts


Who can you trust with your WebRTC service?

Chris KoehnckeChris Koehncke

6a013486861d6a970c0147e044645b970b-800wiWhat happens when your WebRTC service provider goes bye-bye? Tsahi at bloggeek.me maintains an extensive list of companies offering WebRTC services, it’s a long list of far from household names. These companies are fruit flies with their life stages being expand, mate and die.

Sadly, there is no long term story for the bulk of these companies. Their  longevity measured in how long the funding holds out and whether they’ve been acquired. Failing that they hope the business grows like a wild banshee or find themselves nursed along on current cash flow, facing unhappy investors looking for an exit.

GoInstant was a pre-WebRTC company offering a collaboration API  for mobile & desktops (think pusher.com or pubnub), they got lucky with an early take out by salesforce.com nearly 2 years ago. So it’s actually surprising that they’ve announced that 31. August they’ll be ceasing operations. The fact that they’ve continued to support operations for nearly 2 years is amazing for longevity.

The recent acquisition of AddLive by Snapchat is of course worrying  to existing AddLive customers, no amount of assurances or even the strictest of commercial documents is likely to stop the natural order of business.

Using a 3rd party for any portion of your service offering is always risky. Trying to split your business amongst multiple WebRTC vendors nearly impossible at this stage. They’re each so different and the investment in integration remains a work effort.  To some degree, you’ll have to be comfortable with perhaps revisiting this topic more often than you’d like. But the entire reason you did this to begin with was to make your life easier. So how’s your life now?

If communications is a part of your application suite, perhaps you should just bite the bullet and learn how to do this stuff yourself. Master the domain. Own the solution end-to-end. This takes time and money, but at least you’d have the core expertise in-house. However, for many, the relative $$$ spent on a WebRTC service don’t warrant the time to build it yourself, but what are the long term costs?

For WebRTC service vendors, addressing the “what happens if we go bust” is negative selling but is this stopping customers from proceeding, scared that you will indeed go bust or best case get acquired, leaving them in a scramble? How does that chart look in your presentation deck?

Even open source isn’t the answer, nearly everything is built up with layers of technology with most of the value being that you simply got it all to work reliably.

For the moment, it remains risky business, however, if communications will be forever a part of your application, to some degree, your team will need to understand the underpinnings and while risk free is never an option, de-risking your WebRTC powered service element by developing and carefully managing 3rd parties would seem to be the most prudent.

 

Comments 4
  • Nicole
    Posted on

    Nicole Nicole

    Author

    Good points, and I like the fly analogy. Especially agree: If WebRTC is core to your business/application, you should probably learn how to run your own infrastructure.


  • Doron
    Posted on

    Doron Doron

    Author

    Since realistically there is no guarantee for lifelong support from any infrastructure provider, large or small (Google Reader comes to mind, but there are many more examples) and the only thing constant about software is change (whether it’s yours or supplied by a service provider) – it’s now more important than ever to have a backup plan to managed services your solution relies on.

    Case in point – when GoInstant announced cessation of their service, it took PubNub less than a day to offer their customers a painless migration plan: http://www.pubnub.com/blog/pubnub-welcomes-goinstant-developers-migration-tools/


  • Silvia
    Posted on

    Silvia Silvia

    Author

    What you need is an open source library that is provided by a company that is not looking for a quick exit.


  • Travis
    Posted on

    Travis Travis

    Author

    Good post. I think a lot depends on the type of client you’re trying to develop. Webrtc in the browser? Tons of resources. Mobile (native)? Less, but still plentiful. Windows desktop (Winforms, WPF)? Good luck.

    I found the native Jingle to be a mess and worse, not a priority for Google. Becoming an expert and owning the entire stack is great if you have the resources but many teams will go bankrupt while doing so.

    In my case, we went with Frozen Mountain IceLink and have found the product and team there worth the cost.


This site uses Akismet to reduce spam. Learn how your comment data is processed.