Recently a company in UK has announced a service called Babble. It is VoIP of course; the novelty is that they are offering free calls to PSTN in UK, US, Europe, Australia and New Zealand. David Beckemeyer and Richard Stastny wonder how they are able to give away free calls to PSTN. A possible answer can be found in an old entry in Om’s blog regarding a company called StanaPhone. That entry could have been written yesterday by changing the name to Babble. Interestingly when I visited StanaPhone’s website today I noticed that the free offer has been replaced by a tariff structure similar to SkypeOut.
Tom Keating has received his GNUP number. The number is of the form (8844) 1234567890. The initial string 884,denotes a spare country code in E.164 address space. (It is a pity that they couldn’t use 844 to coincide with Peerio444.) It is bad human factors that they are stringing all 10 digits together. They should chunk them into smaller strings. Even MAC addresses and IP addresses are chunked. Just a thought.
Andy points out the recent press release from VoiceWing that announces four new [sic] features: Do Not Disturb, Simultaneous Ring, Permanent Caller ID Block and Forward Voice Mail.
The last one is especially worth noting. Previously they used to offer Voice Mail and Visual Voice Mail (voice mail is displayed in “Personal Account Manager” window where the user can manage the messages as in an email client). Now they have added the ability to forward them to other email addresses. If they are a true “Do not be evil” company they would have allowed their users to view the voice mail in their own preferred email address, instead of forcing them to log into PAM. But then they couldn’t add one new feature today.
Andy points out a recent review of Skype. The original source identifies a couple of surprising information about Skype that I have not seen before:
1. The Skype system suffers from some delays in updating the status of users. While playing around with blocking contacts and going offline for short periods, we found that the status of the user took about five minutes or longer to update on other systems.
2. Unfortunately, there is no 'message waiting' ability as seen in ICQ. Messages sent to an offline user will not be received by them when they log on. Unsurprising given the P2P structure of Skype. They could send the offline message as an email as MSN Messenger does. It is surprising that they do not have this capability.
A couple of weeks back Om posted a quote from an executive of Level 3 wherein he indicated that they are developing a IP phone service including emergency 911 service. Also according to that quote, many dialup ISPs are testing it. Om asked: “Isn’t there a certain bandwidth requirement to make VoIP service work.” I posted a comment suggesting that if Skype can declare that they support dial-up customers why not others also offer a similar capability.
A couple of days later, Phoneboy posted an entry on bandwidth requirement by codec, referencing a paper from Grandstream. I sent an email to Phoneboy pointing out the error in the source and that direct application of those numbers is not appropriate for dial-up connection. Today he posted an update.
Even his update does not satisfactorily answer the possibility of possible Level 3 service. I need to establish the feasibility of using PSTN infrastructure to offer VoIP for my next entry where I discuss SBC’s possible refilling of interconnect tariff.
I pointed out to Phoneboy that iLBC codec uses only 32.8 kbps and so can be used in a dial-up connection that uses V.34bis modem. I also pointed out that introduces 60 ms incremental delay compare to a connection based on broadband. Phoneboy points out in his reply that most of the connection is 28.8 kbps and that dialup connection introduces 150-200 ms delay.
I have no empirical data to question either of these claims. If it is indeed true then I suppose we could use IP and RTP header compression to reduce the overhead on the dial-up link. Since this service will be offered by ISPs, this is possible. I am not sure why dial-up connection introduces 200 ms delay. I suspect that it is because the dial-up ISPs have high contention ratio (as I pointed out in my comment to Om’s entry. If this is so, then the ISPs can reduce the contention ratio for this service. So I continue to maintain that ISPs could replicate “Skype quality” even for dial-up users.
Om comments on a recent column by John Dvorak regarding VoIP. I almost share Om’s opinions but for one. Indeed, my reaction is well depicted in the accompanying picture, where Dvorak is closing his ears in apparent disgust. Dvorak’s points are:
1. Incumbents are hindering introduction of VoIP because of their investment in expensive “old phone switches”.
2. Incumbents are rolling out inferior quality DSL to artificially limit VoIP adoption rate
3. 911 (sic, not E911) service is a red herring. It seems the world was OK when he was a kid and dialing the operator was sufficient. (I presume that the car he used to ride in didn’t have front, side airbags, collapsible and shock absorbing hood etc. Will he be content to do the same now, I wonder.) Also did you know that mobile operators are introducing 911 (I presume E911 is meant here as well) service because of the threat of VoIP.
4. The consumers can avoid the regulation quagmire by using VoIP only for on-net calls and use cell phone to reach those that are still enslaved to PSTN.
Om takes issue with points 1, 2 and 4 (he is silent on 3). I have already betrayed my reaction to point 3. But there is something in point 4 that I want to expand on. Here I am making an assumption that many of us will have access to more than one communication: wireline telephone, wireless telephone and VoIP device; and we will be connected to different set of networks at different time. Recognizing this reality, “VoIP ATA” should have multiple interfaces: FXO for wireline PSTN, “Cellsocket” like adaptor and of course basic VoIP adaptor. So when Om wants to contact Andy, he will select Andy’s name from his “buddy list” prompting the ATA to generate a SIP INVITE which will be intercepted by Andy’s SIP redirect server and it will instruct Om of the current point of contact. On the other hand, if he wants to contact his mother, he would have programmed the “buddy list” to dial out using one of the PSTN interfaces. This ATA can also do further optimization by using some selection logic for using the specific PSTN interface. So, it is possible to adapt the technology to human behavior if we move away from 12+1 button phone interface (my predictable refrain).
The final question in the interview given by Tom Kershaw, vice-president for communications services at Verisign and was published in BusinessWeek Online is worth noting:
Q: So you can do lots of neat things with VoIP. But most people can't even use all the features on their regular phones.
A: That's true, and if we don't find ways to make this compelling to users in the form of new services, we're wasting our time. We might as well not do it. That's one of the key challenges to present this to consumers in ways that they can easily do and understand.
The response to the question why their slice of VoIP business will be a growth business is posted here without a comment.
“What's even more interesting is that in VoIP you can allow customization of those addresses. There's a huge appetite for customization in the under-25 market…Instead of getting an actual phone number, say I want my phone number to be TomK. In VoIP you can do that. It can be TomK on a computer keyboard or on a phone touchpad. As long as it's unique, it works. And you could have different phone numbers that end up at the same phone with special warnings. You could give your mother-in-law a special phone number, for example.”
So it would seem from an article written by two well known researchers from Bell Labs/Lucent and published in ACM Queue. (To be sure that is not their claim, but one could come to that conclusion after reading that paper.) It was recently slashdotted and also widely bookmarked at the community bookmark site de.icio.us. Given the wide visibility given to the article, it is worth taking a closer look at it.
The article repeats the oft stated claim that VoIP will enable new revenue generating services. (Given the anticipated benefits, the resulting apocalypse and the fact that it is fashionable to make fun of the name VoIP, probably we should rename it Kalki.) The article further declares that many engaging services are already available, even though they do not give many examples. Interestingly, Pulver periodically admonishes the industry to come up with some innovative services. One example they talk about is click-to-dial. But it is not clear how VoIP facilitates it. Indeed, the description contained in the paper, the resultant call is a PSTN call.
The article suggests that limited signaling mechanism of PSTN makes it “awkward” to invoke services in PSTN. If so, then the article fails to pint out to VoIP industry that its reliance on ATA, which implies the same limited signaling mechanism, is going to limit VoIP to the same extent as well. The article points out the lack of industry wide effort in addressing the feature interaction problem. Feature interaction is one of the major reasons why PSTN has been slow to introduce new features; the article suggests that this problem is compounded in a distributed environment envisioned for VoIP. But it omits to point out that the limited signaling mechanism further aggravates this issue and this is another reason to revisit ATA based deployments.
One of the issues that stumped ISDN that VoIP is also finding it difficult to address is the transmission of DTMF tones. Whereas ISDN had two means of transporting DTMF (signaling message or on the bearer channel), VoIP provides three ways – encoded along with the speech, RFC 2833 and signaling message. Each method is appropriate for a given scenario, while none is satisfactory for all. The paper touches on this point but fails to point out that a terminal with richer interface can overcome this problem because the user will be able to specify the appropriate mode of transport.
The following sentence is found in their description of SIP: “If one SIP node knows the address of another node, the first may invite the second to join a SIP session.” If so, then how come a whole industry is developing around “helping end-points find one another”? The article does not say.
Ostensibly, the article would like to claim many great things for VoIP. In my opinion it does not make a case for it; it also fails to point out how current terminal architecture is holding back deployment of new and exciting features.
A while back I posted my views on virtual numbers and how one can get it for free from one VoIPUser. It turns out that their claim is a bit misleading (given the reference, it has to be understated). This what they say in their website: “Join as a member and obtain your free UK local-rate telephone number (DID) which can be forwarded to any landline in a number of countries in the World, your SIP software or hardware phone or Asterisk VoIP PBX.” In actuality, they issue numbers from 08 series (and also from 07, which may also suffer from similar problem), which as was pointed out by Geddes recently is unregulated in UK and callers to this number are charged exorbitantly, compared to normal numbers.
One demerit for me for not researching this matter fully.