How many people use Skype? How much? @Skype tweeted yesterday:
“Steve Ballmer announces new stat: Over 200 million avg. monthly connected #Skype users #CES”
“And a 2nd new #Skype stat: More than 300 billion total calling mins annually, with approximately 50% being video calling mins #CES”
The 200 million average monthly connected users for December 2011 is consistent with an end-of-year bump as people substitute Skype for travel. Skype has been running weekly highs of 30+ million and lows of 15+ million concurrently connected for the last six months, more consistently high than in previous seasons.
Skype’s 300+ billion minutes of live talk is a little less than half the time people spend on Facebook, if we go by the 53.5 billion monthly minutes reported by Nielsen for May 2011.
Meanwhile, Skype continues stealing cross-border-calling minutes and hard currency from international telecoms, per Telegeography. PSTN traffic was 438 billion minutes in 2011 compared to their estimate of Skype’s 145 billion minutes; about 1 in 4 cross-border minutes are on Skype. This is up from about 1 in 5 last year.
Roughly half of all Skype minutes cross a national border if we trust these figures.
More than 3.4 million North American Skype users, about 12% of those online at the time, were affected by an ISP service fumble, with reduced or no access to Skype dialtone for up to 90 minutes today. Phyber Communications reports ISPs appear to have been affected by Juniper routers on Level3 networks, including TimeWarnerCable Internet.
A $39.99 box the size of a coffee mug, the FREETALK Connect.me lets Americans plug your dumb phone into your Ethernet network. Your phone rings when someone Skypes you or calls your SkypeIn phone number. You can call out to phone numbers using your Skype credits or your SkypeOut subscription. And you can call Skype users if you add them to a “speed dial” list. Just don’t try to dial 911 or another emergency phone number; Skype will refuse to pass the call through to avoid regulation.
Phone geeks call this an ATA device, short for analog telephone adapter or analogue terminal adapter, connecting an old style phone with a VoIP service. Previous third-party Skype ATAs all died horrible deaths, languishing in warehouses, returned to retailers, or stuck in my old-electronics-waiting-for-recycling box. What makes the Connect.me different?
A few things. First, this is a smarter ATA, not requiring your desktop computer running Skype except to set things up, assign a Skype name, pay credits, or change your speed dial list. Second, setup may be simple, but Skype didn’t send press test units in advance of today’s announcement, so I won’t be able to attest to that.
Some downsides. Freetalk isn’t worried about sound quality; your home phone’s microphone and speakers won’t do justice to Skype wideband audio anyway (G.722, G.729 and Skype NWC codecs). No emergency dialing. It uses an additional power outlet, and is wide enough to require three outlets on a power strip. It occupies one of your router’s few Ethernet ports; no WiFi. It may ring on non-Skype calls. If you use multiple Connect.mes, you’ll have to install the Safari browser on your PC to configure them. You won’t be sending caller ID but you’ll be able to see incoming IDs.
Clearly Hong Kong’s Freetalk, on behalf of Skype, hopes this smarter ATA (SkypeKit inside?) will sell. Freetalk failed to bring its much more expensive Freetalk Connect•All multiline small business phone system to market at $2000. The Connect.me is more their style, cheaper than many Skype headsets.
Skype investors will appreciate Connect.me too, as it spreads Skype’s dialtone so more people are available to make and take Skype calls 24×7. Skype wants your phone to ring when you are away from your home computer or when your mobile phone is off.
User Tip: Dial ** before a number to force the call to Skype. Dial # first to force the call through your local phone company.
From the Skype store FAQ:
What Skype features are supported on the FREETALK® Connect•Me Phone Adapter?
Free, unlimited Skype-to-Skype calls (up to 99 speed dials/Skype contacts)
Calls to landlines and mobiles domestically as well as internationally.
Can I access my Skype address book?
Yes, once you have completed the set up on your computer, your phone adapter will be automatically synced up with your Skype account. So all your Skype contact details and numbers will be ready to use.
Can I make Skype-to-Skype calls to other devices, such as mobiles and Skype-ready TVs?
Yes, you can contact anyone on Skype regardless of the device they are using.
How do I add credit?
Buy your first credit amount online by clicking on ‘buy credit’ at Skype.com. If you select the ‘auto-recharge’ option, Skype will recharge your account automatically when it falls below $2 so you do not unexpectedly run out of credit.
Can I receive calls?
Yes, if someone calls you, your home phone will ring and you can answer as normal.
My landline and router are not close to each other, can I still use the home phone adapter?
Your landline socket and router do need to be near each other as they both connect to the home phone adapter. However, you can purchase a ‘Powerline Adapter kit’ which extends an ethernet network connection to any electrical outlet.
I only want to use my home phone adapter for Skype calls, must I connect my landline?
A landline is not required to use the home phone adapter for Skype. However Skype should not be considered as a replacement landline service.
CORRECTION: eBuyNow Ltd. is in Hong Kong, not Spain.
Every startup founder is getting this question from investors, and customers. This wasn’t true in 2009. The question speaks to two of Skype’s strengths as it approaches its 8th birthday: brand and network effects.
The first strength is brand awareness. Everyone knows Skype. Literally half of Internet users have tried Skype. So Skype is no longer the domain of specialists in telecom, instant messaging, or video conferencing. More people know about Skype than know about Cisco’s telepresence or that Vidyo powers Google+ Hangouts. More people understand you can make cheaper calls on Skype than know of the hundreds of other services that offer even lower rates.
The second strength is network effect. The chance that someone you know is in Skype are vastly better than with any other communication or collaboration service. A user’s social network switching costs are not trivial. You lose history, you lose touch with contacts. You are adopting a weaker dialtone with fewer people you care about available for calls right now.
To be considered, challengers must do what Skype does.
Skype is the new vanilla, the new baseline, the ante for this round. Once you can “skype,” then you must offer something different, something more, something better.
Bonush will try to be Skype voice chat in a browser after it launches. (open for early Beta right now.)
Ooma Mobile was Skype on an iPad, before Skype’s own iPad app.
Vonage’s Time to Call is the voice part of Skype while paying for international calls at Skype rates with pay as you go billing to your iTunes account. Convenient for some.
IsCoord’s is-phone conference for iPad is Skype with SIP without video on an iPad available for white-label OEMs.
Toktumi’s Line2 is Skype with better SMS and telephony features, without video, instant messaging or presence.
FriendCaller is Skype on many devices and in browsers, with a Facebook voice app.
ChatTime is SkypeOut international calling for less money, showing what time it is where you’re calling.
Voxer is Skype without PSTN, adding voice IM and location check-ins.
Apple’s FaceTime is Skype just for Apple and without PSTN service.
Skype still wins. Explicit or not, every time we discuss a product in this space, we invoke Skype.
Rivals (and even the term “rivals” invokes Skype) have four choices:
Do less. Cut features to increase focus, convenience and usability.
Do more. Add features to serve unmet needs.
Do different. Reconceive the problem, delivery, pricing, psychology.
Niche. Serve an underserved market, add insight into a specific context.
Two things complicate matters.
1. Skype is a moving target.
They left an opening for iPad apps for two years, letting others define themselves as Skype+iPad before entering the space. We will see Skype@Microsoft co-brand all sorts of products, from Sharepoint services to gameplay add-ons. This adds danger to filling in a gap in Skype’s product family.
2. Skype is become platform.
Microskype will offer real-time communication components to developers on nearly every Microsoft platform. Mobile, web, desktop, server, you’ll be able to build Skype into whatever you imagine over the next few years. APIs make “do less,” “do more” and “niche” easier for everyone, right down the long tail.
For example, blogs like Skype Journal will offer group video chat for readers of this very page, the site paying a tiny monthly subscription for the feature, free to visitors. It will be part of every “would you like to talk with a customer service agent” widget. Peer-support graphs like WeightWatchers, Quantified Self, and Twelve-Step programs will guide with whom you talk and when.
Platformers like Skype, Voxeo, Tokbox, Jajah and Twilio will power them, commodifying voice and video chat as hundreds of thousands of apps and web sites add realtime talk to their user experience palette.
So what works now? Less, more, different and niche are all viable. You just must be extremely persuasive on why the “better” you offer is worth the customer’s switching costs. Investors will want you to spend toward achieving network effect critical mass.
What works in the long term? Dominating a defined niche (there’s room for only one Grindr) or changing customer expectations, as Skype did to Plain Old Telephone Service.
Someone will change the paradigm, displacing Skype as the iconic reference. Until then, product managers, buyers, investors and the press will ask: how are you different from Skype? Your answer is…
Are you tired of Twilight romance? Of True Blood intrigue (the new season starts this weekend)? Then pick up The Strain. Hogan and Del Toro’s vampire trilogy is a dark, fast, medical thriller; an epidemiological Robin Cook meets Nosferatu’s creepy dread.
We meet The Strain’s vampires at New York’s JFK airport, a superbly evolved viral infection, spreading wildly through populations despite the best efforts of the Centers for Disease Control and Prevention, who are better prepared for a zombie apocalypse. When you are turned, you mindlessly return to your family to spread the vampire contagion. As you mature from a mindless, bloodthirsty revenant, you develop wit, self-control, focus and social organization. Society can’t take you down by killing some master vampire; every bit of your vampiric biological matter flourishes independently, finding niche after niche. Humans, dogs, vermin. The spread of this organism is so fast, wide, and effective, this could be an extinction event for the human species.
This is such a Skype story.
Consulting clients often ask me how Skype grew so well. I say virality was built into the product’s DNA from the start. You need two people to talk. So once you get the Skype bug, you drag your first circle into Skype, likely your family. Then you infect coworkers and friends. Finally entire institutions are in Skype’s thrall.
I was infected in 2003, shortly after Skype launched. I told my friends. I blogged about it. I invited colleagues. I bought webcams and installed Skype for my transcontinental family. Skype had a symbiotic relationship with me. I spread Skype and was rewarded with better calls, lower costs, more convenience, and more independence from phone companies.
Most social software manifests as an outbreak, burnt out quickly. Skype managed to become an epidemic, then a pandemic. Epidemic simulations model an infection’s success on multiple factors. Speed of transmission, the time until a new host becomes infectious, mortality or resistance rates, susceptibility distribution, isolation vs. connection of sub-populations, and reinfection rates.
The same analyses apply to Skype. Time until first call, loyalty rates, sharing rates, usage rates, peak social graph, time to first purchase, social graph diversity (a few strong vs. many weak ties), rates of use on more than one device, percent of time with active Skype dialtone.
Skype’s virality let it spread like a plague. Skype’s p2p tech meant it wouldn’t run out of food (cash) while it spread. Skype is still spreading after having touched nearly half of all Internet users.
The Strain is a compelling page turner. Can’t wait to read the second installment coming out in paper soon.
YouTube started with an asynchronous experience. Millions of files being uploaded, slowly. Prepared for different screens, slowly. Cached in content distribution networks, slowly. Watched on demand.
They’ve had many experiments with streaming live video, perhaps going back further than the October 2009 live-streamed U2 concert. 2011’s Royal Wedding was YouTube’s most-watched live stream. This means they had to upload one stream, instantly. Transcode for different screens, live. And cache and distribute live streams simultaneously across all regions.
Millions of live viewers, so, nicely done.
Is YouTube ready for the next challenge? To turn YouTube into a live video calling, conferencing, and casting service?
I’ve asked video and VoIP professionals about this for two years. Everyone says there are three challenges: addressing, connection and latency. Can YouTube users perform people search efficiently and accurately? Can you connect people promptly, grabbing attention so people answer a call? And can you stream the voice and video with less than a tenth of a second delay, so people don’t notice the lag? Industry people say these are hard, especially latency. No doubt. But I have confidence that Google’s commitment and resources can meet the challenge.
When the Google Voice team nails these problems, they are free to innovate user experience and market applications. To build live conversation into Google properties. To offer live conversation as a platform for AdWords advertisers. To define video as the default Android calling mode. To make your Google identity more important than your phone number.
Where does that leave Skype? Will they launch a cloud Skyping platform before Google? Will it be as compelling for today’s users and developers as the first Skype desktop clients were in the Summer of 2003? I know they aspire to a new degree of awesome.
Yet it probably won’t come down to quality or design. Network effects attract users, so the people you want to talk to or work with are within the network. Network effects trump product quality and user experience. Multiply network effects by the ability to reach people in the network. So can your network offer dialtone all day, everywhere, in every context?
Android gives Google an edge in network dialtone, always on in your pocket. Skype will have to be strategically awesomer to beat that.
And yet I’m unable to see an outage in Skype’s stats showing the number of users online. I can only imagine the service interruption was short, the interruption was intermittent, it affected only a few people, or the data we’re seeing is incorrect.
The last and first quarters of the year show seasonal growth spurts in the number of people signed in to Skype. Just marking when the number of concurrent users crosses a new high-water mark, like it did at 30 million users yesterday, you can see the growth slow as Spring comes to the northern hemisphere.
Concerned that seasonality may become more noticeable…
“Our rapid growth may have overshadowed whatever seasonal factors might have influenced our business to date.”
On seasonality’s effect on profits [emphasis is mine]:
“Our net revenues exhibits seasonality because many of our users reduce their use of our products with the onset of good weather during the Northern Hemisphere’s summer months and our users tend to use our products more in the fourth quarter during the holiday season resulting in weaker net revenue growth during the second and third quarter of the year. Furthermore, we experience significant spikes in the use of our products during significant world events, such as Christmas and the Chinese New Year, or regional events, such as the recent volcanic eruption in Iceland. Due to our high revenue growth rate, the timing of product launches and movements in foreign exchange rates, the seasonality trends are not substantial when reviewing our quarter over quarter results since January 2008.”
I’m only looking at Skype’s dialtone, the people who show to call or be called. Skype hasn’t said much more on other elements of seasonality affecting usage or income. I’m sure investors would want to know:
North/South. Is the seasonal behavior consistent between hemispheres? Does South America slow down as North America speeds up? Are there opportunities for offsetting some seasonality as Skype grows in Africa, southern Asia, and Oceania?
Are other indicators seasonal? Logging in and revenue seem to be seasonal. What about adding contacts, call frequency (do I call more often), call duration (are my calls longer or shorter), call size (are more people in my calls), call mode mix (are more of my calls video)?
Is workplace Skyping also seasonal? Does it vary by industry? By occupation? By country? Where is enterprise seasonality stronger or milder than consumer seasonality?
Does seasonality vary by age? Gender? Income?
Does product commitment moderate or accentuate seasonality? Do active users with many Skype contacts and many conversations demonstrate the same seasonal ups and downs as newbies with just a few close contacts?
Does Skype on mobile phones dampen seasonality? When you don’t need to go to a landline, are you more likely to stay connected to the network? Has Qik experienced the same seasonality over its last five years?
How similar are Skype’s patterns to seasonality experienced by other companies and industries?
Are you North Americans and Europeans slacking off on your Skyping as the weather clears up?
We have a new high: 30 million people logged into Skype at the same time. I’d guess this puts the number of active Skype users about 180 million, since most people turn off Skype at night and don’t log in every day.
Dialtone is one measure of the size and capacity of a network. Your chances of finding the people you want to talk or work with rise as more people join and use the network.
Checking just now, 29,527,474 accounts were connected to the Skype network. That will be close to today’s peak. That’s a lift of 12.8 million from 16.7 logged in last night. This suggests Skype is running at about 177 million active users. Skype’s dial tone is a measure of the value of the network to users; the ability to call and be called.
The last high-water-mark was 17 January 2011, with 28.2 million. That’s a 4.3% bump of 1.3 million in 35 days. This rate of growth seems to be seasonal, based on previous years. We should see growth slow to a crawl by April. Nevertheless, my prediction of reaching 30 million concurrent for this year seems awfully conservative.
Will we see 29 million concurrent Skype users next week as the Lunar New Year continues and people Skype home when they can’t travel home? About 230 million people make the annual trip to their families but hundreds of millions more can only call. These Skype dial tone peaks usually happen on Mondays (in the Americas and EMEA) with the start of the workweek. That this one fell on a Wednesday (Thursday in Asia) signals a change in skyping behavior.
Or will the next peak come the week-after-next for Valentine’s Day, when romantic lovers call longer? Stay tuned.
Last week 28.2 million Skype users were online at the same time, setting a record high water mark for Skype dial tone. This week’s high was 28.4 million, a 0.7% growth in seven days. If this keeps up for half a year, dial tone will rise to 33 million and active users to 200 million. Dial tone is a measure of the capacity of Skype’s network to help people talk with each other.
Contrast this with Tencent QQ, the world’s largest IM network. At the end of September 2010, QQ had 118.7 million peak simultaneous online IM user accounts, chart on the right, about four times Skype’s current dialtone. QQ had 636.6 million active IM user accounts in September.
Two data points aren’t a trend, so let’s see if Skype keeps picking up more new users who sign in more days of the week, stay connected longer each day, and connect using more touch points (mobile phone, office computer, home PC, or home TV).
The number of Skype accounts concurrently logged in to the Skype network hit a new high of 28.2 million today, despite (because of?) this being a national holiday in the United States. Dial tone is a measure of Skype’s capacity: those logged in can use Skype’s services. More importantly, they increase the chances that someone you know is available IM or live voice or video chat.
The “Skype everywhere” product strategy makes this possible. Skypers are signed in from desktops and laptops where Skype is often preinstalled, mobile phones, and televisions.
A hot stranger will IM something inappropriate to you. (NO. Skype’s new filtering technology has been surprisingly good in 2010.)
Skype‘s SilverLakeification will be complete, with a very short leash on strategy and operations at first. (YES)
Skype will serve 125 billion minutes of calls. (DON’T KNOW)
Second Life will serve 20 billion minutes of calls. (DON’T KNOW)
Oprah’s television show will end as scheduled in 2011; lots of Skype calls in 2010 leading to the finale as former guests make cameo appearances, holding out for a spot on Oprah’s last (highest-rated-ever) episode. (NO. Oprah switches to Cisco.)
"The Tyra Banks Show" will end as scheduled in 2010. Nobody will Skype in. (YES)
UK police will allege terrorists use Skype (like everyone else). Parliament will demand the PM bring Skype under control. (NO. Surprisingly quiet on that front after previous lobbying.)
Skype 5.x will offer multiparty video. (YES)
Skype 5.x will offer team features. (NO)
Someone will attend a family funeral via Skype video. And forget they are on camera. (DON’T KNOW)
Skype will release a "naked Skype" public beta. This Skype engine, no user interface, will be free/cheap. Hardware developers will like it; web developers won’t. (NO/YES. SkypeKit released as a publicly-announced invitation-only limited private Beta program. Who’d have thought?)
Skypecasts will still be offline. (YES)
Facebook will add voice to chat. (NO. Despite partnering with Skype on directory services for “finding Skype friends in Facebook.)
Skype for Business will account for ten percent of Skype sales. (DON’T KNOW)
Nortel changes its name to Avaya. Or avice aversa. (NO. Assets sold to Avaya, Ericsson, Genband, Ciena, Kapsch; none keep the Nortel brand.)
The world economy will continue to suck. An American commercial real estate crisis will reinvigorate the Great Recession. Good news for Skype as more people work from home. (YES)
24‘s eighth season will feature Cisco’s new midrange video conferencing. (NO. Still just the highest end Telepresence products.)
Skype won’t offer a "Login with Skype" service. (YES)
Google Talk will add multiparty video with On2 inside, and become a standard part of the Google office suite. (NO)
Skypers post thousands of videos of Skype calls on YouTube, thanks to recording software. Jeremy Hague’s Vodburner outpaces Pamela as the bestselling Skype add in. (DON’T KNOW)
The US student loan crisis ($700 billion outstanding) strains consumer lending. (YES)
Skype starts a post-SIP standards discussion about communications protocols for the 21st century. (NO. No public discussion or progress on the ViPR protocols.)
Avaya will make Skype for SIP the default setting for new switches they sell. (NO. At least not yet.)
Skype manages to get a television commercial on the air. (NO. Waiting for new members of the UK Advertising Standards Authority?)
China’s troubled economy will boost Skype usage when families can’t afford to travel home for the Lunar New Year. (DON’T KNOW)
A team will talk for 200 hours in an uninterrupted Skype-to-Skype call. (NO)
Wi-Fi phones will ship with Skype SILK inside. (YES. See Grandstream.)
Six former Skype employees will become CxOs. (YES)
Someone dies, unable to Skype for emergency help. (DON’T KNOW)
You’ll be able to make iSkype voice calls on Verizon 3G before AT&T 3G. (NO)
100 handsets will run on Google’s Android. (YES)
Skype will release their homemade COTTON video codec, so they don’t have to use the ones from Google‘s On2. Higher quality. Easy, free license. Independence. (NO)
Skype.com still won’t let you log in with OpenID. (YES)
Windows Live Messenger gets a huge boost in new user signups from Bing, Office2010, Office Live, and Windows 7. Microsoft will rock in 2010. (YES)
LG ships a television with Skype inside. (NO)
Mobiletelco 3 ships its third generation Skypephone. (YES)
An angry entertainer tweets to a million followers her PC crashed and lost all her Skype history. So she’s switching to… (NO)
Skype opens a mobile research lab in India. (NO)
Gizmo5 features migrate to Google’s plumbing and Google Voice. (YES)
A Harvard Business Review case will feature a Skype-related issue. (NO)
A Fortune 500 company (not eBay) will provide Skype for Windows for their employees. (NO)
Volunteers phonebank using Skype on behalf of a national EU political candidate’s campaign’s. (NO)
A lobbyist slips a Skype-hostile measure into a US law on behalf of incumbent telcos before Skype can muster opposition. (NO. Weak net neutrality rules didn’t require laws).
The Skype store will sell a netbook with Skype preinstalled (NO).
A court will find Skype guilty in a class action suit related to collecting small sums of money from customers but not offering service or prompt refunds. (NO)
Skype will offer to buy Tokbox for its browser-based video. (DON’T KNOW)
Skype revenue per minute called will continue rising from $0.06 as Skype trunking starts to contribute. (DON’T KNOW)
Skype will top $900 million in revenue. (DON’T KNOW)
Skype will sell small businesses pricing plans making it easy to budget and buy. (NO)
An IETF working group publishes avatar portability protocols. (NO)
23 million people will log in to Skype at the same time. (YES. 25.)
180 million new Skype accounts, about 500k daily. (DON’T KNOW)
Skype loses juicy US government contracts over the TOM-Skype security compromises.You don’t know when someone you’re talking with is using a TOM-Skype client with monitoring software from Chinese security agencies. An audit will show Skype on 500K federal employee computers anyway. (NO)
Skype-like features become generic, included in every communications and collaboration product shown at Demo, TechCrunch50, Telephony Startup Camp and similar product launchpads. (YES)
BT/Ribbit adds video support to its platform for programmers. (NO)
Voicemail to email transcription becomes a standard feature in most markets for mobile and home phone service. (NO. I can’t honestly describe what I get in the email as transcription.)
United Nations rescue and recovery teams standardize on Skype. (NO. But almost!)
Skype sponsors a Festivus site for the public "airing of grievances" and videos of your "Feats of Strength." (NO. Dagnabit.)
Hudson Barton predicts a 2010 peak of 27,695,335 Skype users online, Total "real users" will be 67,596,505. (NO. A touch high on the peak: 25,706,903 actual. Same for Real Users: 61,507,824.)
Hi Phil. I’d like to clear up the node to supernode relationship. Each supernode determines it’s client connection list to be full at 350 active tcp connections. This means a supernode is considered at maximum capacity at roughly ~350 client connections however they will and do handle more client connections up to a degree depending on system variables. ATM there are ~58K supernodes "active" and ~116K "idle". My data during the outage listed clients and supernodes however I truncated the dataset to begin at about ~1.5 MM nodes. The decline is linear with online supernodes and clients and represents a "live" graph of the overlay undergoing "segmentation" due to a failure in the main supernode backbone.
CORRECTIONS:
First, while Cain’s data is correct, I mislabeled it. These are 1.4 million nodes, not supernodes. By 3:46PM Eastern in the US, the Skype network crash had been going on for many hours. So you’re seeing the last leg of the collapse. Cain labeled the data correctly; it was my transcription error.
Second, I should have recognized the error. I was confused by two things. All the other information and videos focused on supernode behavior, so I was myopic. I completely missed out on the time zone differences so I didn’t notice that this data really fit into a different place on the event timeline.
Thanks to several folks at Skype who urged me to dig a bit deeper and to anonymous commenter Chupacabra who wrote “1.4 million supernodes? Thats rubbish!”
My response to Chupacabra:
@Chupacabra, I thought that at first too. And you may be right since nobody at Skype is talking. So let’s do some back of the napkin calculations. In the olden days, say 2005 or so, it was said that a supernode could support between 500 and 1000 active nodes easily. On that fateful day last week, there were about 25 million accounts online when things began to go wrong. Very few people have more than one copy of Skype running, so let’s leave that to rounding error for the moment and say there were 25 million nodes running. So 1000 goes into 25,000,000 twenty-five thousand times. 50K supernodes if we say each supports only 500 nodes. 25 to 50K feel right to me.
Are we dividing into the right population? Skype has about 150 million active users over the course of a few weeks, about six times peak levels. So that would make the number of supernodes about 150K to 300K. Still far short of 1400K.
1.4 million supernodes is not what we’d expect.
So,
- Has the math changed? If Cain’s observations were correct, does a supernode now support, say, just 20 other nodes? That would be a massive drop in efficiency.
- Did Cain observe something else, not a supernode but, perhaps, a node that was capable of becoming a supernode?
- Did Cain only find nodes – not supernodes – that were visible from his vantage point in the network? A large but limited subset?
- Did Cain make a fundamental mistake in data collection and processing?
- Did I report it wrong?
I’d like to test his rig, but his response to a request for an interview:
“Hi, Thanks for the offer but I am not speaking to the press for the foreseeable future. Regards, Julian”
Almost 25 million accounts were online when Skype’s 1.4 million supernodes started leaving the cloud. It took 330 minutes for 98% of the supernodes to go offline, cutting off nearly all Skype users. Researcher Julian Cain set up a UDP probe to look at the Skype network as it crashed last week. At the bottom of the blackout, Cain demonstrated how his reverse engineered Skype client attempted to connect to the network and was rejected, like all the other Skype clients struggling to reconnect.
The chart shows 98% of the Skype supernodes leaving the Skype network over 5.5 hours (data). Cain puts minimum viability for the p2p mesh at about 75K nodes.
Cain shot a short video after the network crash. He is showing his “fully functioning 3rd party Skype peer-to-peer stack during the global overlay outage.”
10:42 PM Eastern, still unable to connect with Skype. He shows traceroutes to Luxembourg supernodes operated by Skype, mostly pooled within the same IP ranges. Here’s a list of hard-coded Skype IP addresses from earlier this year. He shows his own reverse engineered, third-party implementation of a Skype client. “It’s the first ever.” He uses Skype’s login server as a bootstrap supernode. “Here I perform a UDP probe of all of the supernodes that ship with the Skype binary to check responsiveness.” He makes his own SkyLib (a core Skype messaging component) and pinged the Skype mega-supernode. He receives a NACK (negative acknowledgement) from the Skype supernode and connects to the network. The network drops his TCP/IP connection, just like any other Skype client in the outage.
Skype promised to report on its internal post-mortem this week. Let’s hope we get this level of disclosure.
Phil Wolff designs and positions realtime collaboration products for effective people. Phil advises the Personal Data Ecosystem Consortium and is a director of the DataPortability Project. Email editor@skypejournal.com, Skype evanwolf, tweet @evanwolf, G+ or call +1-510-444-8234 to talk with Phil. Skype Journal is independent of Skype.
Chart showing timeline of effect of North American Level 3/TimeWarnerCable Internet outage on 7 November 2011. 
Join the Skype 6.X Text Chat Room