Ideally Vonage should have been deploying off-the-shelf VoIP gear in order to work its way round the patent infringement troubles. But I don’t think any vendor is going to trust the troubled company with the post payment invoice option! And with the cumulative amount of existing fines, Vonage will not be left with much cash to buy new equipment and keep going at the same time. One would have thought that the ASP model of Vonage will make the company collapse under its own weight. But it is collapsing for different reasons altogether.

The simplest way would have been for Vonage to deploy off-the-shelf softswitch solution, specifically the call agent and feature server software. The volume of off-the-shelf call agent and applications software needed to replace Vonage’s proprietary solution will cost the company at least $50 million. However, the service provider gets an opportunity to survive. But if you cannot afford to do so, why not try open source Asterisk which has successfully been deployed by several VoIP service providers.

Despite having about $1 billion knocked off their price, I still think eBay paid far too much for Skype. Skype has been instrumental in establishing VoIP worldwide. There is no doubt about that. But I cannot reconcile with the multi-billion price tag. We have not written too much about Skype in the past. The reason is simple: As a research group, we hardly come across any major VoIP company that regards Skype as significant player. However, just for the sake of those who read too much into the writeups in the trade press, here are a few thoughts.

To start, several reporters tend to compare Skype with Voice-over-Broadband service providers like Vonage and Comcast. This is a mistake. Unfortunately we cannot compare Skype with Vonage. Vonage has a ‘tangible’ phone-to-phone service. At least Vonage has its own termination gateways. A general criticism leveled against Vonage is that the company is an ASP only and does not own infrastructure. Well it does own some infrastructure in form of termination gateways and proprietary call servers. It just does not own the last mile i.e. the access part. That begs the question: if Vonage suffers such criticism, how did an infrastructure-less telecom company such as Skype get such high valuation.

In IP TV, the client side of the camp is divided into those who put a lot of intelligence into the solution and those that use thin client architecture. The former would include Nokia-Siemens, Microsoft and Espial. Vendors like UTStarcom, Kasena, Orca, and Thomson all use thin clients. With thin clients the disadvantage is that you take several seconds to navigate through a few lines of EPG.

With the intelligent client you can browse through much faster. If you look at 14 days of programming, which is 2 weeks of data times 24 hours a day times 100 channels, that is a lot of EPG data. And what consumers really want today is better information and better guidance on what is on offer.

mGinger, a garage startup, introduced solicited mobile advertising services in India based on social networking. In a short span of time the company has near 1 million users whom it pays for reading the campaigns. The services are being offered across the country over all major cellular operators.

Click here to read Part II of the interview.
What sort of response are you getting from the advertisers? Who have you signed up so far?

Since 4 months of our launch we have done around 150 - 200 campaigns. We have 70-80 advertisers signed up so far. Besides this we are getting advertising enquiries from all types of advertisers, from small outlets at the street corner to the big boys like HLL. Big names advertising with mGinger would be Taj Gateway Hotels, Tanishq, ICICI prudential mutual fund, Pritish Nandy Communications, etc.

Is there any particular segment you see more advertisers from?

We are seeing a good traction in segments like Hotels, pubs and education. However there is no single category we are confined to.

Source: iLocus

Like other set-top boxes, IP TV STBs do not sell one per home. Depending upon which country you are talking about, it is between 1 and 3 per home. Global average has been found to be around 1.4 per home. So if there are about 7.5 million IP TV subscribers worldwide, we are talking about a cumulative shipment of 10.5 million IP TV STBs worldwide.

Only Motorola and Amino have shipped over a million units. Other major vendors include UTStarcom, Thomson, Yuxing, and Sagem. And there are a couple dozen smaller vendors (doing brisk business with IP TV STBs though) based in China and Taiwan.

In North America, the average number of IP TV STBs per home is an impressive 2.7.

Click here to read Part 1 of the interview.

Any issues with TRAI? They recently came down hard on the tele-callers?

First of all, TRAI is more concerned about the telemarketers who make voice calls. Text based messaging is not that an issue. Secondly, their recent regulation is for Unsolicited Commercial Communications (UCC). Ours is completely solicited. The user permits us to send messages. So we have no issues with TRAI.

Any issues with service providers? You could at times be competing with them for mobile advertising revenue?

Not at all. Infact most of them that we met are very happy with such a service, because they think we are kind of subsidizing the users' mobile bills. We are getting calls from them to ensure that subscribers on their network do get our messages.

Do you have your own sms servers?

No we go through SMS aggregators.

Verimatrix, the IP TV content security vendor, raised further $5 million in funding taking the total funding raised so far to $20 million. The company has activated over 3 million IP TV set-top boxes that come pre-integrated with Verimatrix software. So if around 10.5 million IP TV STBs have been deployed so far, the vendor can claim around 30% market share in IP TV related content-security / DRM /CAS solutions.

Business model: The product sells as client component integrated with STB vendors and server component deployed by telcos at the headend. Verimatrix pre-integrates the client component with STBs free of charge. It makes money when the telco activates that license from the head-end. Verimatrix mostly sells direct to telcos. However the company also works with OEM partners such as Nokia-Siemens and Comverse. It has a fairly large ecosystem of partners which includes STB vendors, middleware vendors, encoder vendors etc.

(Customers, Sales, Geographic split of licenses, Strengths)

Amino regulatory filings in the UK indicate that the IP TV STB vendor has shipped 1.15 million IP TV STBs on a cumulative basis (as of June 2007). Amino was the market leader till 4Q06 in this segment. However Motorola ramped up quite fast during the first half of 2007. Motorola has shipped more than 2 million IP TV STBs so far.

Amino’s internal data analysis shows that a typical Amino enabled customer installs 1.4 IP STBs. So as of June 2007, Amino would be supporting an estimated 820k live IP TV subscribes worldwide.

I would like to have pbx features loaded onto it for intercom. If that could come pre-loaded with some service provider API that would be preferable. When companies like Google host your telephony service under an ASP model in the future, these APIs would be as easy to integrate with your residential gateway as the present day carrier pre-selection.

Anyway, apart from pbx features and pre-integrated service provider APIs, I would also like the gateway to support WiFi so that those family members with WiFi phones can call from their rooms. Or if there is an incoming call, we can transfer calls with simultaneous ring on fixed extention as well as the WiFi phone extention .... and do other 20 odd things with it. If the family members are able to pull up the buddy lists on their WiFi phones and use presence based applications through the residential gateway that would be even better.

It is usually the telemarketers based in India that have to be careful about the Do-Not-Call registry in the US. However, the Indian telecom regulator, TRAI, has recently taken a swing at this in India as well. TRAI recently came out with UCC (Unsolicited Commercial Communications) regulations whereby service providers have to maintain a Do Not Call (DNC) register. Users are allowed to register their number for which they don't want to receive any telemarketing calls. The telemarketers would be required to check this list before making calls to the users on single or multiple operator networks. This register will be centrally maintained by National Informatics Centre (NIC) India.

What specific challenges will you address as the new CEO of Ditech Networks? What will be the thrust areas?

The biggest challenge we have is to achieve wide deployment of our technology within the tier 1 networks. We already have a lot of traction with the tier 1 service providers. The challenge is to achieve wider deployments, particularly with the tier 1 mobile providers.

What products and geographies will you be pushing more?

For us, the mobile market place is significantly larger than VoIP marketplace. So our primary focus is to sell voice quality solutions into the mobile networks. And we do that worldwide. We do have traction in the area of VoIP but it is a smaller piece of our overall business, and we are for now targeting the North American market for that product line.

VoIP industry has been talking about better-than-TDM voice quality for over a decade now. Why do we still have quality issues with VoIP? What is it that service providers are not doing right?

Texas Instruments, Freescale, STMicroelectroncis, Broadcom and Qualcomm are the top 5 communication chip makers. And there are certain common ingredients in all of them.

1. They all cater to CPEs and end points. Since the cost per chip and the sheer volume of end user end-points is huge, that means better margins and bigger sales.

2. Majority of their chip product portfolio addresses broadband and/or wireless technologies. And both these markets are witnessing tremendous growth.

3. Third factor that makes these companies leaders is that they are able to offer solutions across the whole spectrum of higher end and lower end products. So they are not willing to let any opportunity slip away.

4. All these vendors spend at an average 15-20% of their revenues on R&D. Investing in R&D is the only way forward for maintaining differentiation.

I have come across some VoIP integration scenarios within the Facebook environment. But I don’t like any of those options yet. babyTel offers a PC-to-PC client that requires both users to have the same VoIP client running. This approach has not really worked in the past. The other option is to use Rebtel or Jaxtr type application where a user is allocated a local number in several countries. So if I have friends in 20 different countries, I will have 20 different phone numbers in those countries. A) That is a clumsy way of doing things; B) The phone number that is allocated to me in a specific country does not mean that the caller pays nothing to call me; and C) The web element is usually weeded out of the calling

A better way to do VoIP with Facebook would be to integrate an ENUM database with the Facebook API. So you let the Facebook members opt into this database and register their phone number in the ENUM registry, and then specify who within your group you allow to call you. The click-to-call facility will then appear in form of a url to the permitted caller. The advantage you have is that you do not have to list your phone number on the web but still be able to receive a web-to-phone call. And such a call will be free in several developed countries. The other advantage is that if the Facebook ENUM registry connects to something like Spider, then you open lots of new opportunities.

While Vonage is settling in the US with Sprint for patent infringement, a Bangladeshi cell operator GrameenPhone struggles to raise funds to pay a jaw dropping fine of US$25 million for illegal termination of international calls into Bangladesh. That must be the biggest ever fine for the leaky PBX/switch thing.

What I do not get is why the overseas corresponding operators that sent GrameenPhone the traffic would not be liable to pay part of that fine. They are a partner in ‘crime’ too. If organizations like the ITU are too weak to settle such cross border cases, and if there cannot be a court for international telecom trade, perhaps the overseas carriers that interconnect with your provider should be required to sign transparent interconnect contracts that have local jurisdiction.

Sylantro is all pepped up. You just have to see the number of recent announcements to get a sniff of that. One of the customer events I was keen on getting info about was the recent Sylantro Global Summit. We have been shifting focus from Class 5 switch replacement to hosted Voice-over-Broadband deployments within our VoIP quarterly tracking service. Broadsoft and Sylantro are some of the leading vendors in that space. And both had gone a little quiet over the last year or so.

Seeing a flurry of announcements from Sylantro is a welcome change. It is always good to hear from companies at the forefront of innovation. The vendor just concluded its customer summit which saw several high profile attendees. Broadsoft is about to start its customer event in three days’ time. It seems that the message Broadsoft is pushing through is ‘scale’. That is the impression you get from the press announcement about the event. Sylantro on the other hand tried to push through two main messages: multiple use of its platform (same platform being used by wireless as well as wireline networks) and the web extension capabilities. Through web extensions, Sylantro hopes to ‘Re-invent VoIP’. Sylantro has been promoting for several years its openness to work with web developers. There is obviously a lot of potential there. Potential.

I have recently spoken to a few enterprise FMC vendors and the feeling I have is that there have not been many deployments. Can you tell me something different?

We are deployed at over twenty customers. In fact we have announced a couple of them and we should be able to make at least 2 or 3 more names public soon.

The twenty customers that you mention, typically what sort of verticals do they come under?

The most popular verticals for us thus far are healthcare and education. These two verticals tend to be early adopters of new technologies. The other reason is that these two verticals have been the biggest adopters of WiFi technology till date. And the reason they have adopted WiFi wholeheartedly is that people within these verticals tend to be highly mobile within the campus. These are what we call road warriors that can use FMC to be more productive in their jobs. The value of availability is very high in these organizations. Other areas we are finding traction in are construction, hospitality, and high-tech.

A lot of the FMC vendors do not bother about the WiFi part. What are your thoughts on their approach?

If you follow the IP TV market, you must have read a lot about Microsoft’s prospects in this promising global market. I will list some of the main points that I have come across. The reason Microsoft has had some traction in the market is because around the time when telcos started trialing IP TV some three or four years ago there were not many off the shelf middleware products out in the market. Having an early version of the IP TV middleware did not help Microsoft cement its place though. The telecos who deployed the solution found themselves 2 to 3 years delayed in their deployment plans.

And it is not just the delay, there are scalability problems as well. Numbers from AT&T and Deutsche Telekom in the past have pointed at Microsoft being able to support 600 subscribers per server. That is not exactly carrier grade. Microsoft also experienced problems at BT with their DRM software crashing, bringing down the IP TV service. Among the technical and middle management ranks at these telcos there is no great love for the Microsoft product. Microsoft IP TV deployments there are driven by top management relationships.

I posted this entry on ENUM for Facebook type sites. So that is one creative idea for the weekend. I will post another one later today.

Here is my contribution to the field of education: A website that lets students and teachers (belonging to academic institutions ranging from primary schools to universities) upload their assignment work and theses etc. I guess this would mostly be text files. But you could also have audio and video files especially from the likes of Mass Communication students. Multimedia files could also come from other creative arts students pursuing courses related to music, drama etc.

This could be a social/academic networking type of site, marketed by the institutions as well, and should include other relevant Facebook/Orkut bells and whistles. Result: you develop an eco-system of information sharing among students so that the standard of their assignment work improves. If you assign your student to forecast the mobile telephony minutes, as an example, he/she could consult the site and see whether similar kind of work has been carried out by anyone else in some part of the world; and if so what forecasting methodology they used, what variables they examined etc etc. I think that could go a long way in improving educational standards worldwide. The users of this site would really have a strong competitive edge in education.

White labeling in IP TV is probably the easiest. You have content aggregators who are already used to supplying various different custom packages to service providers. The server side middleware can easily be hosted on the wholesaler’s headend. Subscriber management, billing and other back office systems work could also be carried out by the wholesaler. Perhaps the only thing you require is a brandable EPG or the client interface, which has to have flexible GUI that allows the retailer to design the interface as required. I have seen some vendor demos that offer this functionality.

My guess is however that this model will only attract tier 3 service providers who do not have enough resources for the capex involved. That is what we see in white-labeled VoIP. When some of these tier 2 and tier 3 service providers get their hands on some cash, they will prefer to have their own infrastructure. The white label may also be tried by large operators who would like to get some hands-on experience with IP TV (VoIP examples: the early implementations of SBC with deltathree and BellSouth with Pcket8). Most of all, I think your local departmental store chain would be the prime candidate to offer white labeled IP TV (VoIP example: Tesco in the UK).

1. IP Business Trunking services in the US are mainly provided by CLECs. The traffic mix generated by large enterprises is huge and varied: something a CLEC is not capable of handling. Large enterprise users have shown preference to work with the likes of Global Crossing and Level 3 for traffic termination (which can be classified as a subset of IP business trunking service).

2. Quality of service offered by IP Business Trunking providers has not reached the requisite level of quality

3. In order to subscribe to IP Business Trunking service, most of the large enterprises have to migrate from TDM to IP. This they cannot be bothered to work out especially since they are already using TDM business trunking.

4. Large enterprises have fair amount of international long distance traffic, which is not being provided over managed networks by most of the IP Business Trunking service providers.

5. IP Business Trunking tends to be associated with hosted VoIP. At least that is how the service providers usually market it. Hosted VoIP has never been popular among the large enterprises.

After gobbling up nearly US$40 million in fines from two Bangladeshi operators for using VoIP, it is now the turn of India’s largest cell operator Bharti Airtel to shell out US$13 million in fines for disregarding VoIP related rules in India.

I have in fact lost my track of how VoIP is regulated in India. In April 2002, when the international long distance (ILD) market was deregulated, the use of VoIP was allowed in various amusing flavors. Enterprises could use VoIP in closed user networks ‘between H.323 and SIP end points as long as they did not interconnect with the PSTN’. Carriers were allowed to interconnect with the PSTN as long as it was not a retail VoIP service. So you had carriers like Data Access and VSNL deploy VoIP gateways.

Someone pointed at this useful presentation. Slide 9 compares the progress made by the two UK operators with regard to migration to NGN (see below). I must also add input on BT 21CN progress from Cisco (BT uses Cisco VoIP gateway as part of the project): “Commercial traffic between two major UK cities is progressing well”

• BT 21CN
– Roll out 2008 -2011
– 5500 exchanges.
– 100% of the population.
– 1 exchange in trial.
– Cost £10Bn
– 30M lines
– Ericson soft switch
– Huawei/Fujitsu MSAN.
– Ethernet backhaul

• Carphone Warehouse Networks NGN
– Roll Out 2006-2008
– 1650 exchanges.
– 86% of the population
– 1100 exchanges in service.
– Cost £200M
– 4 M lines.
– Sonus soft switch
– Huawei/Fujitsu MSAN
– Ethernet backhaul.

What is the significance of the word ‘Ruckus’ in the company name?

‘Ruckus’ represents our spirit of making a lot of noise in the market. Most people basically said WiFi is unreliable technology and we are able to make it reliable enough to be used as a utility.

You seem more inclined towards leveraging WiFi for Video.

Yes. Look at iPhones and devices like that. Video is already in our culture, and if there is no cable involved they are going to love it even more.

Glad to hear about Microsoft’s VoIP product launch. About time we saw some fresh perspective brought into this promising market segment. With Microsoft’s vast army of worldwide application developer partners, I hope this product unleashes the real potential of IP. Understandably some PBX vendors will be able to handle this, while others will struggle. Let me take up Cisco’s case. Cisco does not do hardware based IP PBX. But yes it bundles phones with the soft PBX products, and it makes most of its money from those phones. With Microsoft’s entry into VoIP, is that IP Phone business under threat? Obviously not. If I use Microsoft OCS and have to use IP Phones along with it, I will still go to Cisco for the phones. I think the real threat to Cisco comes when we bring in mobility into the PBXs. That will impact Cisco’s IP Phone business for sure.

I would like to rub some more into an already overused analogy of what Microsoft did to the computer market. ‘ ….. OS became independent of the hardware and desktop became independent of the server ….’. Microsoft’s Office Communications Server and things like Asterisk are pure software PBXs and decoupled from hardware. The PC equivalent there would be … what? Will it be IP Phone? Yes, in the short run. But in the long run, I think not. In the long run, the PC equivalent is most likely to be either portable laptop or the mobile phone. We are talking about mobile PBX.

• Total users 246 million, up from 220 million at the end of 2Q07

• Revenue: $98 million, up from $90 million in 2Q07

• SkypeOut minutes (PC-to-Phone): 1.4 billion, up from 1.3 billion during 2Q07

• Call volume on Skype-to-Skype (PC-to-PC) has dropped from 7.1 billion in the previous quarter to 6.1 billion in 3Q07

• Revaluation of Skype down to $2 billion. Search on for new CEO

• Moving forward, focus on expanding feature set, better integration with partner assets, and accelerated investment to induce more user engagement

For a while we are back to the square one with the Skype type businesses. For the second time in VoIP history, the number of downloads and free callers do not matter because a company as creative as eBay has failed to leverage that huge user base toward something out of this world. What matters is how much money Skpe is generating. And the source of revenue is its PC-to-Phone application. So far the customer acquisition cost for Skype has been less than $0.01 per customer. Not any more. eBay is out to sign up as many new Skype customers as possible, expand the paying SkypeOut user base, and in the process scale up the Skype revenues. The deal with MySpace is just one of many such expansion driven partnerships.

The revenue sharing arrangement between the two has not been disclosed. The simplest model is paying MySpace a cut out of the SkypeOut revenues generated by the MySpace IM users. The other model is to pay MySpace a fixed amount per new paying customer that Skype signs up through MySpace IM. I think the latter is a more sensible model for MySpace although in practice I think they have gone for a mix of the two models. Having said that, I doubt if the social networking site would want to trifle with the telecom billing, the CDRs and all that frightening stuff. If my guess is right, I am also curious to know how much Skype will pay MySpace per (paying) customer acquisition.

NGN deployment at telcos (specifically VoIP) has seen many diversions over the last five years. To me the first diversion was Voice-over-Broadband (VoBB). Majority of the industry players had placed bets on switch replacement rather than the hosted offering such as VoBB. So that was diversion number 1. Diversion number 2 has been IP TV deployments. Diversion number 3 would be the disappointment or the lack of excitement around IMS. These diversions have basically knocked out many talented equipment makers.

Web 2.0 (or Voice 2.0 thing within the overall Web 2.0 genre) is understandably and potentially a much bigger diversion than the above three. Two things need to happen for Web 2.0 to assert. Hosted applications model will require acceptance, and the extent of the IP cloud or the reach of public and private IP backbones needs to reach a critical size. The bigger the size of that IP cloud, the larger the volume of communication sessions that stays within IP cloud e.g. calls do not touch PSTN so much.

In both these scenarios, the VoIP hardware that does the conversions between IP and PSTN world (media gateway) and the call agents (softswitches) that most of the time control that conversion, will not be required as most sessions will be IP-to-IP.