Search This Blog

Tuesday, April 24, 2007

Wireless vs. Wired


A colleague had emailed me this simple image a while back. They say sometimes an image can be worth a thousand words. I thought this image powerfully depicts the engineering challenges that lie behind making wireless networks 'work well'.

(I have no idea about the original source of the image, so if someone knows, please email it and I'd be happy to post credits).


Monday, April 23, 2007

IMS deployments - on the rise and around the corner


Are you surprised by the title? Well, that is really how I see things as it stands today. Based on my discussions with most OEMs and ISVs who have themselves been trialing all over the world for a few years now, we are just about at the phase when most trials are getting out into real deployment. Incidentally, incase you think that IMS all-IP ‘live’ deployments have not yet happened, think again. Remember, in 2006, there was an announcement that Wateem Telecom selected Motorola for a wimax deployment ? Well, that network does data and is ready for voice. Not sure if you knew, but that entire network is over an IMS subsystem (yes, you guessed it – WiMAX on its own is not a session specific architecture, while IMS is, so it makes sense to have IMS on WiMAX, huh ?). And yes, I mean R5+, which is all IP for both signaling and media.

Several green field operators are already deploying or have deployed IMS driven networks (we work with many of them), but very few are touting the IMS name around right now, since IMS has been a long used, much abused and somewhat delayed technology. But make no mistake, IMS powered networks are on the rise. Based on my discussions with several players, IMS deployments are on a steep rise, partly also due to ongoing WiMAX trials. It is worthy to note that most of the WiMAX trials today are to do with ‘high speed data’ and not voice. In other words, WiMAX as it stands today, and defined by WiMAX Forum is only at the IP-CAN level – all the blazing speed for an internet connection. But not voice calls, voice features or the other session level services. And please, don’t tell me “why do we need anything? P2P solves everything!”. Well it doesn’t. You need centralized services for technical and non-technical reasons. For example:

a) Technical: How on earth you do reliably implement a voice mail system and retrieve it when a person is offline ? What happens if the peer node who stored your VM is offline when you come online ?
b) Technical: How do you implement feature interaction ? (example A forwards to B to C, but by mistake C sets a forward to A) since in a P2P network no one has an idea of what service the other has provisioned
c) Technical: How do you implement dial-plans?
d) Non-Technical: What do you do if your “spectrum provider” (you know, in a wireless world, someone is paying for expensive spectrum) blocks VoIP ?

And again, this goes back to my original statement I had made in an earlier post – “IMS is the only architecture that is out there today which supports mobile, fixed and nomadic networks” from a session, policy and interworking perspective. Admittedly some of these areas are still ‘in progress’ from a standards perspective, but it’s the farthest out there in terms of maturity. It was also good to know that Verizon has decided to open up A-IMS and push it to standards bodies. A lot of vendors and operators have done good (but till date proprietary) work on enhancing IMS and filling in gaps, so it is good to see it get more public.

And finally, if you are wondering about my post because you hardly hear IMS being talked about today, this is because we are in the stage between “Hype is Over” to “Deployment is reality”. Here is how is describe it to many:

Friday, April 13, 2007

Everyone wants to be a YouTube




My apologies for not posting for a while. Besides my own family needs, I’ve been also travelling a lot the past few weeks and visited the usual shows like VoN, CTIA and others. (In fact, Jeff managed to catch me at his show here – I’m the one with the widest grin).

It is always interesting to see what is the next big thing that is there in everyone’s mind. And it is no surprise when I say it is now, “Media”.

A lot of my good friends have either started their own “Media” companies, want to start their own Media company, have started and have changed their business plan thrice already, or are hobnobbing with content producers and aggregators to see if it makes sense for them to start their own media company.

And I use the term “media” in the loosest possible way, because so do they. What they want to do ranges from “Video Streaming” to “Digital conversion” to “lifestyle content” and what-not.

I spent a full day at the CTIA’s Mobile Billboard event, where folks from AMP’D, MTV, Atlantic Records et al came in and spoke about how they view the next decade of media, and I always like listening to their opinions, since they have been in media for a long time, as opposed to us telecom gear-heads who think media is as simple as building a switch.

And as usual, I have my own opinions and thoughts for those in the ‘build a media company’ as well as the ‘mobile content’ fray:

  • It was interesting to note that in the mobile world, people are used to paying for content while the desktop world is littered with ‘free to use – we make money with ads’ schemes. It was fascinating to hear Bill Stone, CEO of AMP’D talk about how AMP’D ARPU is over $100 these days where people are happily paying for polished episodes to be telecast to their phones (They call it mobi-sodes). An interesting side effect of this is that mobile TV is a really cheap way to get into TV broadcasting without the huge costs. Take AMP’Ds relationship with Comedy Central – lil bush was such a success that Comedy Central signed up with them to broadcast the serial on TV as well. AMP’D did not have to pay the millions required to try and enter through traditional broadcast TV media

  • I am personally petrified with the ‘eyeballs to revenue’ business strategy. And that is simply because I’ve seen that to be the primary cause of the 2000’s dot com burst. If you are thinking of opening up such a company, I’d love to see you have a real business plan. It is just my gut feel that one good way would be not to ignore the mobile world. Instead, try and enter the mobile space first with targetted solutions and then spread to the desktop world, and not the other way around. I just think there are too may strong players in the pure desktop market for VoD, IPTV and similar solutions including Google, Fox, Joost, etc.


  • Building a Video/Audio/TV streaming product is not just about slick interfaces. I'm getting a little tired at companies focussing on slick interfaces and oomph factors at the UI level and focussing less on what matters more - non jittery video feed, good lip synching and good caching, effective CPU utilization while making sure my dual core CPU is not choking itself to death

  • You won’t believe how many people think they need to build their own overlay architecture to provide solid video streaming quality. I don’t think that’s a bad thing, but I am yet to see a good quality overlay network. I’ve used Joost too, and I think it’s quality is no better than many others I've tried. I’d strongly suggest people to look at re-using well established content caching infrastructure like those provided by Akamai and Limelight.

  • User generated vs. professionally generated content: It was interesting to hear that atleast in the mobile space, all the content providers today seem to feel there is little or no market for user-generated amateur content. Instead, they strongly believe people will continue to watch professionally created content and that the mobile medium is unlike the desktop medium where user-generated Youtube content is widespread (but again, would you be willing to pay to watch a video some teen created, and if not, how will anyone except Google actually make money ?)

  • I am really against those who are trying to replicate the ‘desktop’ environment on your mobile. Sorry, but I’ve been long playing with content-adaptations and virtual screens for mobile, where the effort is to make your mobile screen as rich as your desktop. I just don’t buy it. Let’s face it, the form-factor is different, the MMI is different and the needs are therefore different. No matter how hard adaptive browsers try (like Opera Mini), it is impossible to effectively predict what is relevant and what is not. I’d much rather have a clean and intuitive interface for the mobile, and I don’t think it should be the same. Which is why I am a supporter of the .mobi movement. Trying to retro-fit two different MMIs (Man-Machine interfaces) together is painful. That is not to say I am against making mobile interfaces more powerful – I am all for it. But make it more powerful to what is relevant to a mobile MMI. On the same lines, I am a little distressed about the new Google and Yahoo mobile search adaptations. Their focus seems to completely be on 'reduced clicks'. Well, besides 'reduced clicks' the next, and if not more important factor is also 'reduced scrolling'. In the effort to produce maximum information with minimum clicks, both these solutions put in lots of gunk on the screen (if I type "web 2.0" it will not only search on the net, but also local, images and the rest). Not fun.

  • Getting bought out as the only exit strategy is scary. In this new content/media space, very few are looking at scaling. Most seem to be interested in getting around a million ‘eyeballs’ and then are hoping to get bought out by the biggies like Google, Yahoo, AOL, and the rest. First, there is a very small chance that you will actually get bought out, and an even smaller chance that you will get bought out for a profitable amount. This directly ties in to your expected revenue model. If you cannot figure out a mechanism where you will get paid per subscriber, you likely have a hard to win model in the long run.