What stops internet access from being cheaper in India !!!

HailStonE

Skilled
Found this article few days back but was quite busy to post it sooner..


Hard Landing

But what about India? In data provided by to the telecom regulator TRAI, state-owned BSNL pointed out that bandwidth charges for a 2,500 Mbps line have fallen dramatically in the past four years — they are now a fifth of what they were in 2008 (see The Cost of Access). So the 'first mile' costs for India have actually fallen sharply.

But it's another component of costs, paid by telecom operators, and internet service providers in the 'middle mile', which have proved controversial. In March this year, TRAI called for comments on a consultation paper it issued on costs paid by ISPs and telecom operators, for twelve cable landing stations, located at different parts in the country (half of them in Mumbai), through which India connects with the rest of the world.

Such stations are the key points on the map through which submarine cables cross into the Indian mainland, and link up with the domestic telecom network. It is at such stations that domestic operators, who buy bandwidth on an international submarine cable, actually connect their systems to the cable.

Tata Communications and Bharti together have a 93% market share in the cable landing station 'market'. Tata Communications for instance, owns 5 cable landing stations, which together handle 56% of the total 'activated' capacity of bandwidth in the country. Bharti owns two stations, which together handle 37% of such bandwidth.

The key problem: the costs charged by cable landing station owners to other providers to connect to the global network, are high, haven't changed in years, and now account for a significant chunk of bandwidth costs. According to the same BSNL note submitted to TRAI, while bandwidth cost has fallen, access costs paid at cable landing stations to reach the global network in the first place, have remained unchanged in the last five years, and now account for 56% of bandwidth costs.


.jpg



The comparison with similar charges in other countries are stark — such charges in Singapore are around $1,200-$3,875 per year. In India, such charges range from a low of $150,600 per year charged by RCom to connect to its Falcon cable in Mumbai (part of the FLAG network), to a high of $687,200 charged by Airtel to connect to the Europe India Gateway cable, also in Mumbai.

The costs to be paid are determined by TRAI taking into account the costs of cable station owners (it's effectively cost-plus pricing) and other operators have asked TRAI to actually set the rates of such access altogether. "The cable landing access costs should probably be market-based, but subject to a cap set by TRAI," feels the MNC executive, who has earlier been involved with efforts to set up a landing station.

However, current station owners themselves point out that entry is open - anyone with an international long distance licence is free to set up a landing station. Airtel, in its response to the TRAI paper, points out that stations are likely to increase to 16 in the next two years.

"We strongly believe that the current setup is enough to make the market competitive," says Ajay Chitkara, chief executive of global voice and data at Airtel. "There is no need for a system of regulated pricing." He says such charges are about 5-13% of the end-to-end price of bandwidth.

Perhaps the next big wave of innovation in telecom markets will move from the retail end to the wholesale end, pushing bandwidth costs down even further.


.jpg





Source : - What stops internet access from being cheaper - The Economic Times
 
Was thinking yesterday only.. that how in the world can any movie streaming be possible when even buffering an SD highlight package is a damn catastrophe with our (basic) connection speeds.

IT hub of the world.. and the local net crawls.
 
Another thing to consider is almost all the data we Indians access is on outside servers, which costs money which the operators has to pay when buying the Bandwidth.
In America, most of the time, domestic servers are accessed which costs less money.
And many companies wouldn't start server in India (FB,Google for e.g.) because of our government's microscopic capabilities to understand how everything works and policy to remove 'immoral' content.
 
There is no need to compare India with other countries.

Just compare South India to North India, you can see the fine difference !!

Any Hi Speed connection (2Mbps or above) always impose FUP of 75 -100 GB & costs around 1500rs - 2000rs in Delhi/NCR

While more speedier, 4 - 10 Mbps is available with No Limit or 150GB FUP @ just 1000rs - 1200rs.
 
So basically BSNL gets 1Megabit of bandwidth for Rs419 and its Selling 4mbps with Some GB's and 512bps for Rs.1350 . Is it Really fair?
and moreover 512Kbps UL for Rs.750


God Save US From these Stupids!

Correct me if I am wrong here.. So the landing charges per Megabit is 236 Rupees in the wholesale price of Rs419?
 
So basically for a line that transmitting 1mbs continuously every month, the cost even in the inflated scenario is 419+236= 655 rupees.

Now with a contention ratio of 1:4, it costs them 655/4 (or you can say the cost is spread over 4 subscribers. Its written in their terms mostly.)
So around 164 per subscriber.

The line's throughput (capacity) is around 10.8gb at 1mbps per day. or 324gb per month.
So even if the FUP is a generous 100gb. That means they can fit in not one 3 subscribers here. So dividing the cost by 3 again, it comes to 54.5 ~ 55 rupees.


All the calculations above are in case all the subscribers are using it at peak capacity. Which isnt the case most times and hence the contention ratio which with the current usage pattern you can say is 1:12 and not even 1:4. Hence its cheaper for they companies by a few more folds around 20bucks per 1mbps line they sell with FUP and sharing.

Now for the last mile connectivity, and the service, i doubt they require another 950 bucks per month (assuming they charge 1000rs per month for 1mbps line at 100gb FUP.)


No wonder Failtel can turn over profits even after such moronic service. I'd get into this business ASAP. :p
 
Who said failtel is making profit?. They r making big time losses over the past 3 quarters.

I remember some dude trying to setup ISP in India in a big way but went kaput even before he started. His company was all over TE , broad band forums.

Any one knows what happend to him ?
 
Who said failtel is making profit?. They r making big time losses over the past 3 quarters.

I remember some dude trying to setup ISP in India in a big way but went kaput even before he started. His company was all over TE , broad band forums.

Any one knows what happend to him ?

Yeah that was Hayai. I guess he got caught in the never ending government red tape and such things. Possible other ISP providers must have also played some part in stopping him from moving forward.
 
So basically for a line that transmitting 1mbs continuously every month, the cost even in the inflated scenario is 419+236= 655 rupees.

Now with a contention ratio of 1:4, it costs them 655/4 (or you can say the cost is spread over 4 subscribers. Its written in their terms mostly.)
So around 164 per subscriber.

The line's throughput (capacity) is around 10.8gb at 1mbps per day. or 324gb per month.
So even if the FUP is a generous 100gb. That means they can fit in not one 3 subscribers here. So dividing the cost by 3 again, it comes to 54.5 ~ 55 rupees.

any broadband expert can confirm this??

and companies like ACT and BEAM claim their contention ratio is 1:1
 
In America, most of the time, domestic servers are accessed which costs less money.

It's not just that the servers are domestic, it's that the bandwidth in use is PEERING bandwidth and not TRANSIT bandwidth. The same is true in Europe and the rest of Asia.

In India we don't have that luxury because NIXI in it's infinite wisdom charges Rs14/GB as settlement charges, so most of the time ISPs are forced to peer in Singapore (About $1/mbit/month, no settlement charges) which involves sending traffic through the cables and back...

And many companies wouldn't start server in India (FB,Google for e.g.) because of our government's microscopic capabilities to understand how everything works and policy to remove 'immoral' content.

Wrong. Both Google and Facebook (among others) have servers in India. Google has them in Mumbai (Prabhadevi), Chennai and Delhi. Facebook has it's offices in Hyderabad PLUS any ISP with an Akamai box or peering is connected to Facebook.

I can reach Google's network in a whopping 2ms from my home connection. Well, sometimes 3 or 4ms, but still.

So basically for a line that transmitting 1mbs continuously every month, the cost even in the inflated scenario is 419+236= 655 rupees.

Now with a contention ratio of 1:4, it costs them 655/4 (or you can say the cost is spread over 4 subscribers. Its written in their terms mostly.)
So around 164 per subscriber.

The line's throughput (capacity) is around 10.8gb at 1mbps per day. or 324gb per month.
So even if the FUP is a generous 100gb. That means they can fit in not one 3 subscribers here. So dividing the cost by 3 again, it comes to 54.5 ~ 55 rupees.

That calculation is fine if everyone is connected to the network at equal speed with equal cable at equal distances, but because 99% of BSNL's last-mile network (to use them as an example) is ADSL, there are about 200 other variables you have to take in to account. In their defence, it's not as simple as you imagine, but on the other side of that same coin, they don't manage their network as well as an ISP of that size should and there's a lot they could do to improve it.

All the calculations above are in case all the subscribers are using it at peak capacity. Which isnt the case most times and hence the contention ratio which with the current usage pattern you can say is 1:12 and not even 1:4. Hence its cheaper for they companies by a few more folds around 20bucks per 1mbps line they sell with FUP and sharing.

They're not allowed to sell their lines at capacity. Add in another 33% plus transport overheads and calculate that.

Now for the last mile connectivity, and the service, i doubt they require another 950 bucks per month (assuming they charge 1000rs per month for 1mbps line at 100gb FUP.)

Depends where you're talking. BSNL has an obligation to serve the entire country, even unprofitable areas. You're suggesting that to serve someone in Orissa or Bihar or the North East (say 1000-2000km from the landing station) costs them the same as it does to serve someone in Chennai itself. It doesn't. There's a whole lot of backhaul between the two places and that needs to be accounted for, not just in the capital expenditure but in the operational expenses as well.

PLUS there's the added cost for cross-subsidizing the plans, that is to say, they're probably making a loss on their Rs200/month customers, so they have to recoup that from the Rs1000/month+ customers - they need to average our their costs and prices across the whole nation, not just based on the arbitrary figures of the bandwidth bill.

No wonder Failtel can turn over profits even after such moronic service. I'd get into this business ASAP. :p

Don't. I've probably aged 15 years since 2009. Just a warning :p

Also, Airtel, Reliance Communications, MTNL and BSNL are in huge amounts of debt. BSNL is on the verge of collapse if the USOF doesn't save it's bacon.

Yeah that was Hayai. I guess he got caught in the never ending government red tape and such things. Possible other ISP providers must have also played some part in stopping him from moving forward.

That's a bit irresponsible of you. There are quite a few people using our services who might be inclined to disagree with your comment about Hayai going kaput. As it happens, many ISPs (Fivenet, Beam Telecom, CIPL among others) who were sceptical are now interested to know how we're doing it and how.

Only Airtel themselves seem to be still sceptical (as shown in a recent article on Medianama) but they don't care so long as I pay my bandwidth bills on time... but with Airtel's new FTTH plans it seems they're responding to our offerings.

Government red-tape was a problem for a while but now it's less of an issue than the last mile - that's the real painful bit at the moment.

- - - Updated - - -

any broadband expert can confirm this??

and companies like ACT and BEAM claim their contention ratio is 1:1

No they don't - the Fair Usage paragraph on http://beamtele.com/index.php?option=com_content&view=article&id=149&Itemid=91 specifically states that end-user bandwidth is shared but they don't give the numbers anymore.

According to a telephone conversation I had with one of the DGMs at Beam a couple of weeks ago, they have +/-200 STMs (that's around 30Gbit/s of bandwidth) and with some 3 lakh customers, that suggests a contention ratio of around 1:10.
 
^Nailed it.

PS: I too had dreams of opening a data centre since childhood (and I'll if I get that much cash & skill).
 
^Nailed it.

PS: I too had dreams of opening a data centre since childhood (and I'll if I get that much cash & skill).

Good luck. Hope you have quite a few crores at your disposal. Whatever you think it's going to cost, triple it :)

- - - Updated - - -

Just realized how sarcastic that sounds. But seriously, good luck - India needs lots and lots of data centres IMO. When you need peering, give me a call :)

- - - Updated - - -

Also, I have a problem with this threads title. Broadband in India *IS* very cheap (!!!! I know I know, but hear me out).

What it is not is good.

The question we should be asking is why is Internet in India not better, or to phrase it like the thread title "What stops Internet access from being better value in India!!!"

See, I think that the ISPs such as BSNL, Airtel, MTNL etc would do very well both for themselves and their customers to simply remove speed limits from their ADSL plans (I target the ADSL players because they make up over 80% of the market).

With the existing usage limits in place, they wouldn't need to buy much more bandwidth (if any), and the benefit for them is that it would result in a less congested network. The benefit for you is that you can download & watch stuff better (Youtube videos stream more smoothly, anything you download is finished faster, thereby "freeing up the lines" for everyone else and so on)... I've written about it at great length on other sites/forums so I won't repeat myself here.

Oh, and, they wouldn't need to actually charge you any more money for the speed... the cost of a Gigabyte at 1mbps is roughly the same as one at 16-20mbps (on ADSL)... or in my case (Metro Ethernet and FTTH) 1mbps to 100mbps... the equipment etc is all the same. Same cables. Same everything.
 
^Thanks for the heads-up @mgcarley. I am not a rich lad loaded with crores, just playing in thousands now :p But you never know, world is so small and life is so twisting with no limits towards the sky!! :)

Regarding QoS, I agree completely except in case of Airtel (and Beam ?). Its good at least from what I have used, but pings aren't that good. The big players don't have last mile problem, still they play dirty game. Give what you take O' ISPs! Metro ethenet is still in beta (or alpha) stage, I believe. Because, last mile connectivity along with good QoS is much more important than sucking 10mbps connection.

PS: I really want to buy the name "Ctrl+S", damn it they they already created it. Once I called their CC giving an offer to buy them :p
 
Last edited by a moderator:
^Thanks for the heads-up mgcarley. I am not a rich lad loaded with crores, just playing in thousands now :p But you never know, world is so small and life is so twisting with no limits towards the sky!! :)

Anytime.

Regarding QoS, I agree completely except in case of Airtel (and Beam ?). Its good at least from what I have used, but pings aren't that good. The big players don't have last mile problem, still they play dirty game. Give what you take O' ISPs! Metro ethenet is still in beta (or alpha) stage, I believe. Because, last mile connectivity along with good QoS is much more important than sucking 10mbps connection.

1. Airtel is (according to many and despite a spate of recent "anti-consumer" moves - according to customers, not me) still the best ISP in India as far as quality/stability is concerned.

2. Pings vary from network to network - from what I've seen, Beam's pings (even outside of India) are half decent.

3. Which "big players" don't have a last mile problem? Every time I've talked to MTNL, BSNL or any private player it's *ALWAYS* their largest problem - and one of the largest expenses. Moreso than bandwidth. And what "dirty game" do you refer to? In my case it's not an ISP problem, it's a right-of-way problem - and I know for a fact I'm not alone because even MTNL is facing this issue in Mumbai (which is where most of my experience lies at this point).

4. Metro Ethernet is way not anywhere near Alpha/Beta stage. It's extremely well established. As is what we use (GPON). Both have many millions of connections worldwide. ADSL is still king, but check back after a few years and I doubt it'll have the same margin as it does today.

5. The point of high-speed broadband is to make it so that you *can't* saturate the connection all the time - that is to say, make it sufficiently fast that you'd run out of hard-drive space well before you'd be able to keep the line in use all the time.
 
1. Airtel is (according to many and despite a spate of recent "anti-consumer" moves - according to customers, not me) still the best ISP in India as far as quality/stability is concerned.

2. Pings vary from network to network - from what I've seen, Beam's pings (even outside of India) are half decent.

3. Which "big players" don't have a last mile problem? Every time I've talked to MTNL, BSNL or any private player it's *ALWAYS* their largest problem - and one of the largest expenses. Moreso than bandwidth. And what "dirty game" do you refer to? In my case it's not an ISP problem, it's a right-of-way problem - and I know for a fact I'm not alone because even MTNL is facing this issue in Mumbai (which is where most of my experience lies at this point).

4. Metro Ethernet is way not anywhere near Alpha/Beta stage. It's extremely well established. As is what we use (GPON). Both have many millions of connections worldwide. ADSL is still king, but check back after a few years and I doubt it'll have the same margin as it does today.

5. The point of high-speed broadband is to make it so that you *can't* saturate the connection all the time - that is to say, make it sufficiently fast that you'd run out of hard-drive space well before you'd be able to keep the line in use all the time.

1) Recently Airtel has made a policy to not give connections to students as they don't pay bills, even though majority of its customers are students (At least from what I've seen). That's why I too am facing the problem to get a connection. I got it a few days ago, and Airtel surveyors are daily calling me to verify my "job". :(

2) Beam's & Act's pings are one of the best in India as far as I have seen in India. Its as low as 1 ms.

3) Whatever last-mile coverage is the current status, I believe majority of usage chunk lies there only. Deep penetration (read Laaaast mile) is obviously a dream, and unarguably a big problem. If the bandwidth isn't a problem, why are they implementing crappy FUPs? Many ISPs (such as Tikona & Reliance) are big time suckers. Be it the QoS, or service or whatever, they suck. From dirty game, I meant FUP and hidden clauses (mostly related to capping of speed & bandwidth) which many ISPs implement many times.

4) I said in reference to India, and its reach to the people. Most of them haven't even heard about it.

5) Then 100mbps should be the ideal speed. :p
 
@mgcarley

I never meant to blame Hayai for anything. In fact i was stating that You and Hayai in general must have got caught in the government red tape. In fact i followed your ISP's section on IBF very closely and was hoping that Hayai came to my home soon in Mumbai. I know it is general tendency to blame government and corruption and i admit i do it but that seemed to be the most likely case at least with your ISP (strictly my opinion).

I completely agree with you that point of high speed broadband is so that i do not saturate the bandwidth. Take my case here in US. In India i had a 256 Kbps UL connection which meant downloading a 700 MB took several hours. And here in US i have a 15 Mbps FiOS connection and it takes hardly few minutes to get the same task done and i have freed up the line after my download is finished. This is something most of the ISPs in India do not understand. They think that if they give 100 Mbps we are going to download stuff at 100 Mbps full day which is not going to happen because that would mean we will need TBs of storage at our disposal.
I am totally in for a high speed broadband with reasonable FUP like 100 GB.
Your plans were (are) bang on target and they are what is actually needed rather than money milking plans of likes of Failtel.
 
Back
Top