Very many people have claimed that one of the attractive aspects of Skype is that “it just works” behind almost all kinds of NATs and Firewalls. It was revealed more than a year ago that the scheme Skype uses is a standard UDP hole punching scheme and that it is no different than what others do using the “dreaded” Session Border Controllers. This was later confirmed by a more scholarly article. Not withstanding this, majority of articles treated as if Skype uses a secret technique. For example, as late as last month Bill Campbell insisted that since Skype has not revealed their method and since Skype claims that the supernodes do not consume more than 5 kbps, relay nodes are not used. Now Skype has released an official booklet where they describe their NAT traversal scheme, which is none other than UDP hole punching (pages 3 and 4). Now that it is settled, let me repeat an implication that could affect Skype.
It is well know that if the NAT/Firewall is not symmetric, then one can optimize and eliminate the relay node after initial exchange. I am sure Skype uses this and that is how it reduces the load on the supernodes and the relay nodes. But if the NAT/Firewall is symmetric then a relay node must be present in the direction of the traffic terminating at hat NAT/Firewall, for the duration of the session. This increases the amount of bandwidth consumed by the relay node. Since many claim that they just leave a Skype session on forever and given the fact that Skype does not do voice activity detection (to maintain the NAT/FW binding), eventually Skype users will device ways to avoid becoming a relay node. Of course a simple way is to place the Skype client behind a symmetric NAT. If this trend catches on, then Skype has to deploy their own relay nodes, breaking their cost structure. By the way, Skype’s plan to embed clients in the home routers is one way to counter this possibility.
One of my recent entries received unusual attention from high places. Richard Stastny called it a requiem for SIP. How can it be? It didn’t follow the usual norms; there were no platitudes to the dead; it contained bitter remarks. Instead, my entry was intended to be a cautionary note to Skypers that one day they also can be discarded by using the same sword – “bellheaded”.
More than a year back I have expressed my opinion about Skype and I have not changed my mind. But others may have. For some it has even become "iPod of VoIP". I thought I will make it clear by posting a comment to Richard’s entry. He responded, not with a reply comment but a new entry. He tells us that he knew it all along (but then why call my entry to be a requiem to SIP?) and then goes on to say that he wanted to provoke SIP community to change their wayward direction. No matter, the “requiem” label has stuck. So let me state the following in the clear instead of using allegory:
And one more thing. I just plain do not agree that SIP is dead or even need to die, just as surely H.323 need not to have been killed. So please do not call my post to be a requiem.
Andy Abramson quotes an AP story on Yahoo!BB phone service. According to that story, 80% of VoIP users are in Japan. We all know that that can not be true. After all, Skype has 35 million registered users and Yahoo!BB has only 4.5 million. More seriously though, Yahoo!BB phone service is different than the garden variety VoIP service providers and they are also different from Skype.
First of all, Yahoo!BB phone is not a replacement for the traditional service. All the subscribers get POTS from the incumbent, DSL service from Yahoo!BB. The phone service comes with BB service whether you want it or not. That is why the subscriber number is high. This caught even some bloggers in the VON forum by surprise, The DSL modem has built-in ATA with some smarts: it fails over to POTS under failure conditions, emergency calls are forwarded to the POTS line. But all other calls (it seems that in Japan even local calls are metered) are switched over to Yahoo!BB network before being terminated at the PSTN, at a reduced tariff. In other words it is an unadulterated arbitrage play. “Mind you, there is nothing wrong with that.” You will never catch them saying that PSTN is a dinosaur, unlike some of the VoIP evangelists. They are also not Skype because the only way you get the phone service is if you subscribe to BB service.
But they share something with Skype. Since their tariffs are usage based, the subscriber count is not consequential, except for mC2 marketing. So the really useful metric is ARPU. As far as I know Skype’s number is not publicly known. But Yahoo!BB phone numbers are known – about 1100 yen per month. Interestingly it is almost the same amount that Skype is targeting.
This entry uses the poem The Bull by Ralph Hodgson. I learnt this poem in my high school and has influenced my over time. The lessons apply to all of us most of the time. It applies to VoIP players as well, as I argue in here. By the way, my daughter helped me to locate the link. Even though both of us used the same tool, Google, she beat me while I was only struggling. Another instance of The Bull.
Some six years back, I was in a meeting with one of the central figures of SIP. He was talking about the upcoming revolution facilitated by intelligence at the end points and this marvel of a protocol called SIP. Since I was (still am) smarting from a decade old wounds inflicted during ISDN days, I wanted to caution him. He will have none of it. Indeed he derisively welcomed me to the capitalistic world. (Forget that I probably have lived in that world longer than his conscious life time.) So I remembered The Bull:
With his mother gaunt and lean
In the valley warm and green,
Full of baby wonderment,
Blinking out of silly eyes
At a hundred mysteries
... he came to sultan power,
And they owned him master-horn,
Chiefest bull of all among
Bulls and cows a thousand strong.
And in all the tramping herd
Not a bull that barred his way,
Not a cow that said him nay,
Not a bull or cow that erred
In the furnace of his look
Dared a second, worse rebuke
Pity him, this dupe of dream,
Leader of the herd again
Only in his daft old brain,
Once again the bull supreme
And bull enough to bear the part
Only in his tameless heart.
And so it goes…
Martin Geddes observes that “[Skype] hasn’t taken the feature set very far past the PSTN.” He goes on to suggest that currently Skype can’t do much beyond what one can do with all-you-can-eat PSTN plan and an IM client. Indeed I have been saying for a long time that this statement is true in general and not just for the present time.
So once again, there is no hope of new features and arbitrage business is temporal. What do we do?
Russell Shaw recognized that the incumbents could mount a good resistance to the onslaught of VoIP features. Excepting, this was discussed last September and again last month. I also relish in pointing out that these are made possible by Intelligent Network technology.
But there is one trick of VoIP that PSTN can not possibly duplicate – autonomous communication between two IP end points.
Frequently the blogosphere will be peppered with a report claiming that one or another country has declared that VoIP is to be illegal. This of course will be accompanied by a sanctimonious claim that these regulators don’t understand that IP knows no boundary. Such stories take leave only when port blocking needs to be written about. Today, I came across one such entry that has compiled a list of countries that have made it VoIP illegal. Since I know that some of the claims are not true, I did my own research for other claims. (The entry does not provide links to any of them, save one.)
It turns out many of them are “tabloid claims”. For example, the entry claims that “South Africa - Warns citizens that making calls using VoIP will be considered breaking the law.” I got the impression that it is the South African government that has issued the warning. But according to The Register, it is SA telco Telekom that THINKS network bypass is illegal. So first it is not the government; second it is not VoIP, but bypass that could be illegal. The distinction is if the call is end-to-end IP, then legality is not questioned. I have written many times in here as well as comments at other places to suggest that in India it is perfectly legal to use VoIP as long as you do not bypass international gateway. But it doesn’t matter. Some of the other entries are simply too much. It seems that in Qatar, ONE “user reported not being able to use Skype from either a dialup or his ADSL connection while voice over MSN was still working.” Turkey is “apparently not happy about VoIP but tolerates it for now.” And so on down the list.
What is my position on regulating VoIP? Each country has to decide for itself on the regulation. But if somebody were to ask my opinion, I would tell them to make sure that the regulations are consistent. For example, allowing free data communications, but blocking on-net VoIP sessions is pointless. Alternatively, if they allow free PSTN interconnection for VoIP, then they should be prepared to all traffic to masquerade as VoIP to take advantage of the arbitrage scheme. Don’t let “voice” in VoIP color your thinking. Finally, don’t pay attention to the claims that there are no borders in the IP world. Germany successfully enforces its anti-Nazi regulations and all search engines abide by them. Many ultra high-speed users in Hong Kong have different limits on bandwidth consumption for domestic and international end points.
My request to fellow bloggers is to be more precise when they report on VoIP regulations. There is a difference between a ban on on-net VoIP calls and requiring tariff while interconnecting to PSTN. Also licensing regime for service providers is not tantamount to banning VoIP, when a taxi driver is required to get a license; but if this is used to harass a potential competitor of an incumbent monopoly, then for heaven’s sake let us join in the outcry.
Some ten years back, I joined this humongous, multi-billion dollar product development effort. One particular marketing person had a simple way to justify the project. He would start with the worldwide telecom business market which is like hundred or so billion dollar. Surely we will be able to corner this miniscule fraction of a percent of this huge market. How can one disagree with this modest claim? But the resulting revenue opportunity is enough to justify the development cost. Why did I think of this today?
James Enck wanted to guess Skype’s worth based on its revenue. It turns out Rodrigo A. Sepulveda Schulz has a rough guide on estimating Skype’s revenue potential, which comes to $130 million.
As Einstein told us long time back since C is sufficiently large, m could be atomically small.
Richard Stastny informs us about opening ENUM database to the numbers in the range +43 780. This is intended for convergent services that assign SIP or H323 URI. The database can be accessed from both Internet and PSTN. The system allows for the end-users to change the NAPTRs thus "porting" from one provider to another.
I am curious to know Skype’s reaction regarding the URI requirement. I suppose Skype can allocate a SIP URI and then map it to Skype address. But then this scheme allows for easy porting from one service provider to another; probably the anticipated Skype “lock-in” is not so assured.
VoIP has this schizophrenic tendency that periodically sprouts itself. Though we were assured that features never before imagined will be available with VoIP, cheaper price due to efficient IP networking, compared to the aging and clunky PSTN, was pushed by the industry. But in the last few days there has been a ping-pong is worth noting.
Let me start with Mark Evans’ note a couple of days back. He tells us about a poll commissioned by SunRocket suggests that consumers do not have a full understanding of VoIP and he suggests that this is because the industry is not educating the consumers on the benefits of VoIP and cautions that if the industry is focused on price alone, the industry will stagnate.
Irwin Lazar disagrees. Based on his discussions, he feels that people think VoIP is a low cost replacement of PSTN and that they don’t want any fancy features. But worse yet, when they start to use VoIP, they will conclude that VoIP does not fully replace PSTN. He cautions that the resulting negative backlash is not good for the industry.
Martin Geddes uses three advertisements he found in a travel magazine to lament the fixation among service providers to offer arbitraged phone service and “[n]o new value, no new features, no unique convenience, no innovative distribution or pricing or sales method.”
As if in response to Martin’s colorful dig at the industry, VoicePulse took a press release out claiming that they will offer a $50 reward to current and former customers of competing VoIP services if they will switch over to them. I am assuming that PSTN customers are not eligible. Have they concluded, as Mark warned, that VOIP adoption has reached a plateau?
With that let me add my take on AOL’s Callways service and the recent upgrade to MSN IM. To be consistent with the spirit of this note, I will not comment on the cost of AOL’s service. As you probably know already, it is ATA based service offering augmented by additional user control through client applications. Just like many other VoIP offerings, the subscription comes with a telephone number; but unlike some of them, the allocated number will be conforming to the NANP allocation to the area where the service will be used – meaning no “virtual numbers”. I was surprised to hear this. After all Powell told us that this ability to have any number is a great breakthrough. Andy suggests that this restriction is related to E911 service they offer. Yes, they do offer E911 service, IF you do not access the service from another location. If this is the real reason, then it is a lame one; why not allow subscribers to get virtual numbers as secondary numbers?
Two additional features that AOL talks about are Dashboard and Call Alert. Subscribers can mange their account and call handling logic via the Dashboard. Conceptually it looks similar to what many other service providers offer. There might be some subtle differences in the UI; since I have not used it, I am not qualified to comment on it. Call Alert seems to be an extension of their earlier service with the same name that they offered to their dialup customers. For me this looks very similar to Verizon’s iobi service (offered to PSTN subscribers). Given that they offer an ATA, Ithey are tacitly assuming that users will not be close to the PC all the time. So won’t it be nice if the Call Alert information is presented to the phone as well. I hope they address this in their future releases.
In the recent upgrade, MSN is focusing on end-to-end voice and video communication. In anticipation of the increased traffic, they have streamlined the NAT/FW traversal technique and made it more efficient. This means they do not have to deploy too many relay points and consume lots of bandwidth. Other than that I am not sure what is new. I am a user of MSN IM; but still I can not tell.
Almost a year after its initial ruling, CRTC has spelled out its ruling on offering 911 service to VoIP subscribers. The ruling is bound to be hailed widely, but in my opinion this was a lost opportunity to set an innovative solution that we as an industry should strive to achieve.
CRTC has ruled that those service providers who require their customers to access the service from a fixed location (dubbed “fixed” VoIP service) should offer the same level of 911 service as the incumbent would have offered in that location. To understand the nuance, we have to understand that depending on the location, the incumbents offer either E911 or basic service. If the service provider allows the subscriber to access the service from any place (dubbed “nomadic”), then the service provider is required to offer only basic 911 service. In either case, the service must be operational in 90 days and the subscribers must be notified of the availability and the level of service both in the marketing material, service contract and starter kits. Also, warning stickers must be provided that could be affixed to the telephone sets.
They have also identified a third category of service providers, called “foreign exchange” service providers, who “allow(s) users in one exchange to receive telephone calls dialled as local calls in another exchange that they have selected (e.g. a customer located in Ottawa with a Halifax local telephone number).” The ruling states that these service providers must offer basic 911 service within 90 days. In my opinion, this is an unnecessary classification. Either this foreign exchange service is a primary service or a secondary service. If it is secondary service then for 911 ruling this service is immaterial because this is an incoming call service and 911 is an outgoing call service. If, on the other hand, it is a primary service, then the other two classifications must be used. For example, if I use a Halifax number in Ottawa, but my service provider allows me to access service from a single physical location, then I should get E911 and not basic service.
I feel that most of the VoIP service providers will be able to meet this regulation, because they are already able to support basic 911 service requirement.
This is all well and good, but a bit disappointing in that they should have also set a far reaching objective. I feel thay should have stated that as a goal, the industry should develop a method by which applications can talk to the “physical layer” to get location id that it can use it in its communication to the 911 agency. To take a specific case, the ATA must be able to query the DSL modem and get an ID that it can report it in the session initiation towards the 911 agency. The agency can in turn consult its database to get the physical location, just like it does today in PSTN. Once this is in place, even “nomadic” users can get E911 service. I understand this capability is not available currently; but proddings from agencies like CRTC will help set the direction for the industry.
Starting a couple of months back, the buzz was that Google is working on VoIP. It was started with the flimsiest of evidence and subsequently any and all information was used to support the thesis. Om tried valiantly to put a break. Now probably we may have a more concrete information.
After reviewing Google’s 10-K, Mark Evans tells us that Google will spend about $500 million in capital expenditure focusing on “internal projects such as video, desktop search, comparative shopping, blogs, e-mail and localization that complement its search technology.” There is no mention about telecom related expenditures. So at best, they could only be working on development. Either that or VoIP is so inexpensive to deploy that it doesn't show. Right?
Recently I expressed that Net Freedom means that traffic flows should be allowed based on the network resources the flows consume without taking into consideration the specific applications. These entries were referenced by a few other bloggers. I am prompted to elaborate on this simple point based on some of the comments.
Om gently reminds me that as I dream about open and fair net, I have to pay attention to profit making. I most assuredly agree with the need for the service provider to make a decent profit. The service provider can realize any level of profit; for that matter they can charge even usury rates. This is not incongruent to the goals of Net Freedom. “I am paying for this microphone, Mr. Green”.
While writing about ClearWire’s story, Irwin Lazar points out to a discussion in the mailing list populated by North American Network Operators. That discussion suggests that ClearWire may be concerned about processing the large number of packets voice traffic generates more so than the actual bandwidth it consumes. This very well may be true; that is why I used the generic phrase “network resource” rather than “bandwidth” in my Net Freedom credo. Accordingly ClearWire should specify a threshold for pps generated by a user and ban traffic only if a flow exceeds this specified threshold; it shouldn’t matter which application exceeds the threshold.
One of the participants says that VoIP will generate 100pps and that it is too much for the radio system to handle. But this assumes the standard usage scenario of generating a packet every 20 msec. The real advantage of IP Communications is that there is no one set way of doing things. A session can start in one mode but can easily change it at any time as long as the two ends agree to it. A ClearWire customer can suggest use of Push-To-Talk to abide by the pps restriction. Assuming 90kbps media flow and 1500 bytes MTU, the session will generate only 8 pps. Surely this must be perfectly acceptable to ClearWire. This exemplifies why a service provider can not infer the nature of the session and so should not make decisions based on that inference.
To summarize, according to Net Freedom, the Internet access service provider is allowed to charge any amount for the service and allowed to place any restriction at the Network layer or below (assuming the “Invisible hand” is there to take care of abnormal behavior). But beyond that they are not allowed to do any additional filtering or redirection.