The Indian Economy Blog

February 23, 2007

What Is It About Bandwidth in India?

Filed under: Business,Infrastructure — Reuben Abraham @ 11:17 pm

There is no shortage of hype about India’s superpowerdom in IT and Telecom. However, the truth seems to be far from it. Why is it that there is not a single provider out there who is capable of providing decent broadband access, for instance? The prevalent logic seems to be that the only thing that matters is price, seemingly oblivious to the fact that some consumers may be price-insensitive when it comes to fast and reliable broadband access. The U.S. is not the model to follow, but even so, you could get 3-4 MBps thoroughput at any given time for $40 a month (Rs 1800 approx). I bet there’s at least a million households in India that are willing to pay Rs 2000 a month to get decent access. I certainly would be willing to pay that much.

Instead, all you have are providers who claim to provide you high bandwidth, but then place download limits. What do you need 1 MBps for, if you can only download 1 GB worth? A single album from Itunes is about 700 MB, so what’s the point? Surely, you don’t need high bandwidth to check emails? If they give you unlimited download, they restrict your speed to 256 KBps maximum. So, what the hell are these guys thinking? Or is there some problem with the bandwidth available to each of these ISP’s? It seems to make no sense otherwise, when there is an obvious business model in price discrimination, right?

I know the available bandwidth in India today exceeds 20 TB. The question is how much of this is lit. Does anyone know? Secondly, are there any ISP’s out there that are actually providing high bandwidth residential access with unlimited downloads? I am sure lots of IEB readers would like to know, besides me.


  1. I for one would definitely love to know even though I am not in India. Every time I visit India, I give up the most mundane of the tasks – checking email which hardly requires bandwidth. Download speeds are retarded slow on what is supposed to be high speed internet connection.

    Comment by Dhaval Shrimankar — February 24, 2007 @ 3:52 am

  2. Reuben,

    International bandwidth (the 20Tbps that you mention) is only part of the story. There’s plenty more domestic capacity which can be used if the content is hosted within India. That can bring down the retail price of broadband even lower.

    How much of the 20Tbps is lit? A teeny-weeny fraction. First of all, not all the 20TB is available. Design capacity of submarine cables is the ultimate capacity it can achieve over its lifetime (approx 20-25 years). To make use of Moore’s law, the lighting up of the cable (essentially adding electronics at both ends of the cable) takes place closer to demand. Right now, even equipped capacity could be in the range of a Tbps. Utilization, would be less than that.

    The good news is that the 20 Tbps is not all in the hands of one player. VSNL, Reliance and Bharti share this almost equally and they are building more. So expect international bandwidth prices to decline over time.

    My own take is that if you are looking to bring down broadband prices (or increase bandwidth provisioning to homes), you must focus on increasing PC penetration. India’s PC penetration is abysmally low. If PCs take off even at a fraction of the rate at which mobile phones are taking off (or like cable TV did) then bandwidth/prices will not be an issue.

    Comment by Nitin — February 24, 2007 @ 9:05 am

  3. Nitin, thanks for the extremely informative comment. I got another back-channel e-mail that said the lit portion was an extremely small percentage. In any case, at least the available capacity is good news, since I remember a time when New York city had more bandwidth available than the whole of India. Then again, NYC south of 59th street had more phone lines than the whole of India too at one point :)

    However, I am not sure about this point of yours:

    “My own take is that if you are looking to bring down broadband prices (or increase bandwidth provisioning to homes), you must focus on increasing PC penetration. India’s PC penetration is abysmally low. If PCs take off even at a fraction of the rate at which mobile phones are taking off (or like cable TV did) then bandwidth/prices will not be an issue.”

    I don’t know if there is a chicken-and-egg situation here. Would South Korea, for instance, have an equally high PC penetration if they did *not* have a fantastic broadband network in place as well? Does always-on (with the relevant content) high bandwidth drive people to buy PC’s at home, for instance? I am not sure I’d be that keen to do much at home on a laptop if I only had a dodgy dial-up connection at home.

    Secondly, even assuming you’re right, there must be a huge market out there that is willing to pay for the fast broadband, right? My own guess is that there are 1 million households out there, but chances are the numbers are higher. Why doesn’t any ISP want to cater to that market, especially since they’re likely to be present in certain clusters and not in others? Do the economics not work out?

    Comment by Reuben Abraham — February 24, 2007 @ 11:58 am

  4. The working logic behind current tariff plans seems to be that primary users of internet, even at home, are professionals / business users (actually heard it from one of senior telecom executives). That being the case, very well satisfies the logic of 10GB limits.


    For me, though i’m a pretty heavy user, a 256 kbps pipe (if it actually delivers this speed) with a 10GB limit (always expandable) serves most purposes – from plain old office email to watching videos on Youtube. What disappoints me about the current setup is the total lack of network reliability. Within the last 2 years, I’ve used 3 providers – one PSU, one private biggie and, the current one, from a small cable operator. And none have failed to disappoint me – connection goes off at will, not too often but not rare either. Worst of all, none of them have ever been able to give me a timeline of when it’d be back. For businesses that claim to service business needs, i don’t get it how they can afford to be so lackasaidal about reliability!

    Ofcourse the murphy loves me a lot – my connection never fails while blogging or while listening music / watching movies. 5/5 failures happened when I had some urgent office work :(

    Comment by raven — February 24, 2007 @ 4:15 pm

  5. “For me, though i’m a pretty heavy user, a 256 kbps pipe (if it actually delivers this speed) with a 10GB limit (always expandable) serves most purposes – from plain old office email to watching videos on Youtube.”

    In which case, you aren’t that heavy a user, raven :) A heavy user torrents files and downloads albums. One hi-res movie is about 2-4 GB, so a 10 GB limit will not do :) What’s more, a 2 GB movie will take hours to download at 256 KBps (which, as you point out, you never get anyways).

    Comment by Reuben Abraham — February 24, 2007 @ 6:35 pm

  6. I could be wrong, but I believe this has to do with insufficient demand. On one hand you have some people willing and able to pay higher sums for unlimited high-speed b/band access, and on the other hand internet penetration/ broadband penetration is abysmal.

    Maybe my data is ancetodal, but I cannot help feeling that there is insufficient demand and not enough people see the advantages of accessing information, entertainment, education on the net; or of e-mail. I even walk into homes that have internet access that are happy with 500 MB download limits and who use their connections for no more than a couple of hours a day or not even that.

    Part of the issue has to do with a lack of an e-commerce ecosystem, while some part has to do with simple lack of awareness of the power of the net (yes that is true for a large part of the population). I bet if you could get people excited enough about these you would have huge pent up demand that would send the right ‘price/ volume’ signals to ISPs and they will get their act together is finding a way to offer unlimited, hi-speed access.

    Comment by Little Ram — February 24, 2007 @ 10:12 pm

  7. Secondly, are there any ISP’s out there that are actually providing high bandwidth residential access with unlimited downloads?

    Tata Indicom has a 512 Kbps unlimited plan for Rs 1,850 per month.

    That is the max speed for an unlimited plan for residential access, as far as I know.

    Comment by Sudhir Chukkapalli — February 25, 2007 @ 10:00 am

  8. The countries which have good broadband access (5-30 MBps) like South Korea, Singapore, Sweden have built those networks at the expense of taxpayers. Those countries which have really good broadband access (50-100 Mbps) like Japan have been able to build their networks at the expense of paying customers, but only because those customers live in dense areas. So, unless the government is willing to fork over some money to build network infrastructure (let’s say in some form of grants to only those localities which show interest), I’d say most people living in dense cities will have an enormous advantage over those living in rural areas.

    Beyond the dense/rural penetration rate jargon which remains a problem even in the US, it’s really up to an ISP who actually owns the physical network (as opposed to one leasing parts of it) to offer more bandwidth at a reasonable cost. The UK used to have these virtual ISP’s which leased bandwidth capacity from the physical network owners and placed stringent bandwidth limits for customers because their FCC mandated that the owners let everyone have access to their physical network. It was a mess, and only got better because BT decided to actually compete with it’s network. ISP’s are stingy, they don’t want inceased usage that might make them invest in higher capacity. The economics are very similar to a utility, which are better off being public I’m afraid.

    There are no really good ways of building up network infrastructure, unless you happen to live in a dense city (in which case it isn’t a problem). If you’re going to go by the South Korea example though, you would have to build rural networks at the expense of taxpayer’s money.

    Comment by Lot W — February 25, 2007 @ 11:06 am

  9. Reuben,

    I agree that broadband can drive up PC penetration as well as the other way round. My comment was specifically with respect to the current market situation in India. If you are a marketing chap at an Indian ISP, you will look at your addressable market, and that is constrained by PC penetration. As I pointed out, there is sufficient bandwidth supply—both domestic and international—, and there is competition. Then why are the ISPs content on trying to sell it as a premium service to a small market, while they can, actually sell it on a wider scale.

    (I’m discounting the effect of billing and collection issues here)

    Comment by Nitin — February 25, 2007 @ 3:53 pm

  10. Even if full 20 Tbps is fully lit, it can serve only 10M people @2Mbps in a country of 1000M people. Unless the overall capacity is doubled or tripled, it can’t meet the future demand.

    I’m stationed in Singapore and have access at 6Mbps for as low as Rs.1600, I guess 2 Mbps is fair enough for a smooth browsing.

    Comment by harsha — February 26, 2007 @ 1:53 pm

  11. 2MBps or 2Mbps? 2MBps is a lot of speed and not specifically necessary for an indiviual user. I remember during my college days at IIT Kanpur when the whole institute had to share two 2MBps lines!
    Still I think there is a sufficiently big market for 2Mbps customers. I get around 0.5 Mbps for just Rs 700 with unlimited downloads. It’s not that bad after all!

    Comment by Jodhbir Sachdeva — February 26, 2007 @ 3:42 pm

  12. Building public infrastructure for broadband access, I would argue, has immense social externalities, that far outweigh any burden it would impose on the taxpayer. The applications of broadband access for a billion are mind-boggling.

    Comment by Abhishek Nair — February 27, 2007 @ 1:16 am

  13. Building public infrastructure for broadband access, I would argue, has immense social externalities, that far outweigh any burden it would impose on the taxpayer. The applications of broadband access for a billion are mind-boggling.

    Yes! That’s exactly what South Korea, Singapore, et al have determined and they’ve been proven right because of the “web 2.0″ companies to come out of those countries. If the service sector economy is to continue it’s robust growth, this has to be a high on a list of priorities.

    Comment by Devang — February 27, 2007 @ 3:45 am

  14. I am fairly happy with my download speed but i always seem to hit the limit even after avoiding large files.I think increasing the download limit is more important than the download speed.

    Comment by Tushar — February 27, 2007 @ 3:17 pm

  15. The situation is indeed strange. It is almost impossible to get *any* broad band access (At least in Bangalore) today from any provider. They all seem to have hit capacity issues. BSNL will start providing connections to the backlogged applications from April, AFAIK. Airtel and Tata indicom refuse connection in many localities due to capacity bottlenecks.

    I wonder if the key is to find easier/better ways to get the copper into our homes. Or maybe, wireless. WiMax, anyone?

    Comment by tantrik-porter — February 27, 2007 @ 5:55 pm

  16. Just like wireless growth outshadowed wireline subscriber growth in Telecom industry in India, if we wait for a year or two, Wimax or similar technology should eliminate the need for DSL/Cable connections. besides who wants dug up streets in already overcrowded streets in our metros.

    Billions of dollars worth copper based telco infrastructure is getting obsolete in developed world due to wireless and VoIP technologies. In a way, that was money saved for developing world.

    Lastly, the estimate of million Indians willing to spend Rs.2000 per month on broadband seems way off the mark. Internet access to most in this country means horoscopes, orkut,, matrimonials, and bollywood searches. Indians like to spend on tangible things (Roti, Kapdaa, Makaan + cell and car these days)…visit any mall or market and you will realize.

    Comment by Ashutosh — February 27, 2007 @ 9:25 pm

  17. Sudhir,
    I called Tata Indicom and I got the amazingly clueless customer service ever, so I decided to leave that be until the Tata guys themselves are sure what services they’re providing.

    “I get around 0.5 Mbps for just Rs 700 with unlimited downloads. It’s not that bad after all!”

    Jodhbir, who is your provider? Tata Indicom? Who else out there provides 500K without download limits?

    Tantrik, Ashutosh,
    I met a couple of guys in the past who promised me Wi-Max access within a week’s time. Needless to say, nothing has happened since. I think people have a tendency to talk big about the services they’re just about to provide, but the service takes eons to show up.

    Comment by Reuben Abraham — February 27, 2007 @ 9:56 pm

  18. WiMax isn’t all it’s built up to be, especially if the backbone isn’t sufficient. Don’t expect a large amount of end-user capacity increases unless all the pieces within the network (ie. backbone, last-mile connection, etc…) are densely built. I don’t know how good Airtel, Reliance or Tata’s backbones are but I can’t imagine them having as much capacity as Sprint might in the US. What you might ask is the main advantage Sprint gets for having built a backbone which has more aggregate capacity? Simply being able to advertise their wireless CDMA broadband service to sign up new customers, as opposed to Cingular’s service which places an untold limit on monthly downloading and offers worse speeds for a service which costs as much as Sprint’s does. Maybe Reliances’ backbone is similarly built to Sprint’s since they both use CDMA. It probably is considering Reliance plans on making telecommunications free in the future. Let’s hope they don’t leave out the rural areas.

    There are people who have dedicated their lives, if not made it a cause in the US advocating for a faster end-to-end internet network. There are many real world examples of what works and what doesn’t in different conditions along with plenty of ideas on how to make it better.

    Comment by Devang — February 28, 2007 @ 1:03 am

  19. There is something that India can do with respect to increasing broadband availability and ensuring quality.

    President Franklin Roosevelt built up the telephone network and extended it throughout America through a federal program. The result was that the whole of America was connected and the people were able to take advantage of the multiplier effect.

    The Indian government could do something like that. The problem is the large size of the public debt and the high rate of consumption in the economy. We just don’t have enough savings to invest.

    Comment by yum yum — February 28, 2007 @ 4:53 am

  20. WiMax or similar wireless technology with a broad range holds the best promise in dense urban for adoption in india, especially in tackling the last-mile issue. But as with any of the other options, it would need lots of investment apart from obtaining the license from the government for the use of radio frequency needed for setting up a WiMax. I’m waiting for Reliance or Tata companies to announce this kind of investment, I’m sure they will get the return on their money with all the demand.

    Comment by Kiran — February 28, 2007 @ 5:32 am

  21. Kiran,
    If you want to offer decent 1Mbits service (which isn’t oversold) to 1,000,000 people living in a 2500sq. km area, it’s cheaper to lay down wire in a checkered grid for the 50km x 50km city rather than put up enough wireless access points (each one providing ~70Mbits of total bandwidth in the case of WiMax) which will serve the densely populated area. UWB might be a little more suitable in this case when it comes out of the labs and when spectrum is made available for it. Even then, a wired network can’t be beat for dense areas–even when talking about initial costs. It’s also safe to say that a wired network is much more easily maintained.

    On the other hand, if you want to serve a 500,000 people living over a 250,000 sq. km area, each with a 1Mbit connection, offering wireless service to the end-user -with- a fully wireless backbone mesh might be a good option, saving you the trouble of having to dig up anything. The users are more spread out, meaning each access point will have enough bandwidth to handle end-user traffic and traffic just passing by from other access points.

    You might actually need the same amount of wireless access points in either case (because in the later you’re doing everything wirelessly), in the earlier case though, you might as well use WiFi because of the densely populated area. The newest version of WiFi offers a max total bandwidth of 540Mbits/AP compared with 70Mbits /AP for WiMax–but with the increased range. Compare both of those with wired connections: a 1Gbit switch with 5-ports offers a max total bandwidth of 5Gbits, orders of magnitude greater with some caveats. Look at total network capacity per $. WiMax AP’s don’t cost $50 like WiFi AP’s and Gigabit switches do.

    If we’re talking about large numbers of people using high bandwidth applications, wireless will only work in certain cases, otherwise you’ll end up with a slow network making everyone unhappy. The numbers in my examples are half made up, and half from talking to a friend who is working for Sprint to roll out WiMax service in the US. They’re accurate enough is what I’m saying. :)

    Comment by Devang — February 28, 2007 @ 8:18 am

  22. Here’s my experience: I have a 256 Kbps connection from Bharthi/ Airtel at home that costs Rs 1100 (including service tax at 12.2%) per month. This gives me unlimited downloads. I’ve had a bad connection twice, both in the middle of the night – both times, a 24 hour helpdesk resolved the problem within 5 minutes.

    There are umpteen broadband providers in the country. 95% of these buy bandwidth from Tier 1 ISPs like Bharthi/ Reliance/ Tata (VSNL) and resell them as broadband connections. To make more profits, they multiplex more than permitted connections onto one pipe.

    This means: If a local ISP feeds a 1 Mbps connection into my residential block, he can technically sell only four 256 Kbps connections. But he will sell about six of these under the assumption that (a) not all the users will use it at all times and (b) even if for some period of time they do use it concurrently, simple browsing (which is what most homes use it for) will not get drastically affected.

    For a reliable connection and good QoS – I recommend Airtel. The Reliance support service and billing systems are not reliable and Tata: – the less said about them for homes, the better.

    Comment by Koreth — March 3, 2007 @ 12:25 pm

  23. Australia planning big spend on broadband

    Govt. investment is the answer.

    Comment by yum yum — March 5, 2007 @ 3:21 pm

  24. absolutely agree with reuben, i myself gets frustrated with bandwidth and money paid for it. even today in india Consumer is at Mercy of service provider just like small shop guys looting there customers for big profits, most of big businesses in india are doing same.

    infrastructure in almost all areas in india is old and crumbling, to a great extent it stinks for sure.

    hope theres a change soon and we get retail chains sooner backhome so people understand why its good to do shopping for less money.

    Comment by vijay kurhade — March 5, 2007 @ 3:43 pm

  25. Absolutely valid point you have raised.

    I have even seen ads in Mumbai call 64KBPS as lightning fast broadband.
    Now is there anything more to add..?


    Comment by Gaurav — March 6, 2007 @ 6:55 am

  26. Airtel provides by far the best bandwidth in India. Go for their unlimited download plan (Rs. 1200/- per month), think it’s available in Bangalore in select places only so far.

    Comment by Shankar — March 6, 2007 @ 7:21 pm

  27. A heavy user torrents files and downloads albums. One hi-res movie is about 2-4 GB, so a 10 GB limit will not do :)

    Correct. One of my friends downloads over 80 GB a month. Internet is always on at his home. And yup! downloading takes a painfully long time of 30 days with the computer being switched on for 24/7. On the other hand, we have bsnl offering a 2 MBPS broadband and the limit of 2.5 GB sinks within a day or two which is so damn bad. They should offer atleast 20 GB. 256 KBPS is not broadband speed. It is just a faster dialup connection. But considering where we were a few years and where we are now, the improvement is phenomenol. I certainly would be willing to pay 2K for a quality *real* broadband access.

    Comment by Aswin Anand — March 7, 2007 @ 3:06 pm

  28. I realize we aren’t talking about the US, but local ownership of the last mile could be the a part of the broadband solution everywhere: Five Ways Public Ownership Solves the U.S. Broadband Problem. Anyone know how happy Reliance or Airtel might be if local governments started operating on their turf?

    Comment by Lot W — March 8, 2007 @ 1:53 pm

  29. The problem of lack of bandwidth can be attributed to the following set of factors (1) There is no responsible politician who is tech savvy like Algore Ex.V.P of USA, Mr. Maran and co are doing just gimmick. (2) There is no consorted national effort to harness the human capital in this country. Our political leaders do not pep up the enterprising elements but they implicitly dampen them.(3) Unless there is a popular demand as the demand for the colour TV in T.N.State Election no body would hear you.(3) The legal frame work in this county especially the laws governing the contracts and their implementation must be throughly revised. Still we are following the Contract Law based on premises built in previous centuries by the East India Company. The free portal and Email service contracts as done by Yahoo and MSN are void in our country. Even then the urge to make digital document at lowest strata of the society must be heard from all the sections of this society. Certainly the Babus of Government who are more concerned about their service benefit than their productivity and self denying their own tech interest, will not let any such popular demand being transformed into meaningful action. So this is merely a clattering shop.

    Comment by jayavel — March 9, 2007 @ 1:01 pm

  30. People are never satisfied with what they have. Only four years back I was on 128kbps without download limits and I thought that was lightning speed considering the speed that I had on dial-up. My pc was on 24 7 downloading movies and stuffs but now that I have 4 mbps with unlimited download (though there are speeds of upto 24mbps with expensive subscription), I still think that its not fast enough as it takes ages to download dvds. Now, that is because I am in UK but considering the frustration trying to talk over VOIP services to my brother back home in Nepal who uses 256 kbps shared bandwidth broadband and pays much more than what I pay here, you guys are very lucky in India. As there has been a deal between Nepal Telecom with one of the internet gateway in India, I am expecting the service to get better in coming future.
    The message I am trying to convey here is that we never seem to get enough. Even though India s consumers might get more bandwidth at reasonable rate soon, the evolution of next generation high definition dvds with 45 to 50 gb capacity and high definition IPTVs will make the increase in bandwidth lagging far behind the requirement. So I think we should be content with what we have got for the time being and wait and expect that these techie guys come up with genius ideas to increase the internet speed exponentially year on year.

    Comment by Savvy — March 12, 2007 @ 5:50 pm

  31. Nitin makes a really good point about the low PC penetration. I see your argument that if broadband internet was offered, RELIABLE broadband that is, then there’d be more PCs, but that would still be a guess. A company would have to offer top tier broadband with the CHANCE that more people will take advantage of it. Obviously companies would want to profit, and you said it yourself, why wouldn’t they want to make money? There has to be someone on payroll position telling the higher ups that there isn’t a demand in India for anything more than 1 MBps. India is indeed a huge market though, just off its sheer population size alone, your estimates of at least 1 million customers isn’t far fetched. Just tough it out for now. I don’t think anyone doubts that the broadband speed will get better.

    Get Valuable information about T1 lines, T3 lines and DSL lines.

    Comment by Vic — March 25, 2007 @ 8:05 am

  32. Just wanted to say I use Star broadband in Delhi which, according to the little box in the bottom right hand corner, states I get 100.0 mbps (is that good? I’m computer illiterate) with unlimited upload/download for rs. 500. It’s good enough for me, though I often have problems with it’s reliability. I cannot fault the cable co. though, since my computer is very old (great grandpa) and had enough viruses over the years to fill a pond (sigh- I have lots of guests who don’t care what they download). I can’t watch videos etc though, but don’t know if it’s due to my half-dead Win. 98 computer or the cable guy. Someone might be interested in said cheap connection, eh? Star uses 24 online client (whatever that is) which is what I log into. Hope that proves helpful.

    Comment by User — April 18, 2007 @ 12:42 pm

  33. [...] Slow Speed Problem This blog says india has 20 TB of bandwidth. And that was 1 year ago The Indian Economy Blog What Is It About Bandwidth in India? Still not sure how much total bandwidth is avaialable in [...]

    Pingback by Airtel Broadband Slow Speed Problem - Page 10 - India Broadband Forum — February 1, 2008 @ 2:11 pm

  34. As per the discussions going on here i agree about the services being provided by ISPs in INDIA.I work with one of the indian ISP co.What we have seen is that comapnies have their own strategy for their benifits then think of bigger profits which could be gained from corporates so its very obvious that households are being neglected.And more thing is that now after the rules being strictened by TRAI companies know about each other and are trying to woo the customers now by promising better service but i believe that is possible but will take big time.

    Comment by raj — April 15, 2008 @ 4:08 pm

  35. My good you guys in India really have a crap deal when it comes to broadband access. Paying Rs 1100 for a 256 kbps connection is daylight robbery. Even in Malaysia where the wired broadband market is monopolised by one company I only pay about Rs 800 for 1.0 Mbps unlimited and I think even that’s a rip off. What kind of salary do you have to be earning to afford what seems to be still a luxury. People In UK can get 2.0 Mbps for under 30 or 40 pounds these days I think. Back when I was there 1.0 Mbps was about over 35 pounds in 2003. I think India’s middle class has fairly poor purchasing power relative to their earning power and your economy has very high inflation rates. Even though you’ve got a strong IT sector you need policymakers with the vision to push the future development of affordable high speed access particularly in the dense metros first because the payoffs are good here and then start focusing on the more rural areas. I think the Indian government after having profited from economic growth needs to seriously invest in infrastructure development in a big way and I think they should be taking on a very green and environmentally friendly approach to development to the extent that some of the most sophisticated countries in Europe are doing.

    Comment by suren — April 26, 2008 @ 4:46 pm

RSS feed for comments on this post. TrackBack URL

Leave a comment

Powered by WP Hashcash

Powered by WordPress