Another update: Skype stabilized
I wanted to continue to update you on the status of the outage. We've been able to successfully stabilize Skype due to the dedicated supernodes deployed by Skype's engineering team. We're at roughly 90 percent of normal user volumes. — Audio, video and IM are running normally.
Context & Ripple Effects
This is day two of Skype's worst outage since the August 2007 failure, when the company posted hourly updates at midnight GMT while a peer-to-peer bug took the network down. The difference this time is the fix: rather than waiting for the P2P mesh to heal itself, Skype's engineers deployed their own dedicated supernodes — an admission that a network built on other people's computers needs a fallback it controls.
The recovery post follows yesterday's outage notice, and the story has travelled well enough that the L.A. Times is picking up the CEO's suggestion that paying customers may receive credit vouchers. It lands awkwardly for a company that had spent the autumn on offense — a Windows release with Facebook integration and group video calling in October, an Android app before that, and confirmed preparation of new web-based products.
First-order effects
- Paying Skype customers are the immediate constituency: with volumes still around 90 percent of normal and the CEO floating credit vouchers per the L.A. Times pickup, the company's first revenue-side test is whether compensation matches the disruption paid subscribers experienced.
- Skype's engineering team has effectively re-architected the live network mid-crisis by standing up dedicated supernodes, trading the cost savings of a volunteer peer-to-peer topology for control over availability.
Second-order effects
- A second major outage in just over three years gives enterprise and business users — the audience behind Skype's 2007 workplace deals — grounds to weigh redundancy or rivals for calls that matter, putting pressure on Skype's free-first positioning exactly when it is pushing paid features like group video calling.
- The dedicated-supernode fix sets a precedent inside Skype: once engineering maintains fallback infrastructure, the case strengthens for running more of the service on owned capacity, which collides with the economics that made the P2P model cheap.
Third-order effects
- If the pattern holds across both outages — 2007 and now 2010 — Skype's defining architectural bet, a closed walled garden running on users' own machines, becomes a liability it must progressively backfill with company-operated infrastructure, accelerating the confirmed shift toward web-based products away from the native desktop client.
- For the wider VoIP market, each high-profile Skype outage reframes reliability as a product feature rather than an assumption, favoring services that own their infrastructure end to end.
The trend: Consumer peer-to-peer services are learning that availability they don't fully control eventually forces them to buy back that control with dedicated infrastructure — one costly outage at a time.