I live in Portugal, and here the ISPs charge separately for national and international bandwidth. Portugal is a country with 10 million people, so you can imagine how much Portuguese Internet that is. 90% of the traffic I consume is international, oh yes, maybe that’s why the Internet is for: to search for information that is not available to you right by your side.

In the picture there is an interrogation Google or Yahoo should pay more because they consume more bandwidth. I thought that they paid their own servers and bandwidth already… I also pay for the bandwidth when I use Google or Yahoo. The search engines are giving me organized information so I can get access to the most updated information on any subject possible.

If the search engines were to pay more money for their bandwidth, and if that became a considerable cost they would have to start optimizing for lower bandwidth costs. This would inevitably mean less organized information, and less updated indexes of the Internet.

What do you guys think?

What I meant was paying for what is advertised (the promise of bandwidth) and then having to deal with the reality (the absence of bandwidth), which is another. I have no problems with, as you put it, paying for bandwidth AND usage.

However, well known ISPs that operate 3G HSPA or Wimax connections on this model (and include a FUP that outlines usage policies with no mention of packet discrimination) STILL don’t deliver QoS to match what’s advertised and throttle some network traffic to boot. Comments elsewhere on this forum attest to this, and so does personal experience. A more robust answer will have to wait until your results late October, but I’m not sure whether you are looking at packet discrimination as a specific issue? If so, what’s the tool you are using?

That the situation is similar in the UK is small comfort – a bit like saying the human rights situation in not that bad in Sri Lanka because, you know, the US exercises extraordinary rendition, carries out water-boarding and maintains Guantanamo too.

“Packages where the users pay by byte – may be above a ceiling – and may not necessarily be proportionately” is fine with me. It’s not the only way to manage traffic, but it’s arguably what gives the most incentive to Sri Lankan ISP to invest in better infrastructure and connectivity for all. My fundamental question to Rohan was whether and how the economic models of broadband proposed by him / you and Lirneasia (which I do not have a problem with) have a positive bearing on packet discrimination, and what evidence there is for this.

Rohan for example advocates carte blanche, but does not go on to define, “mild forms” of traffic discrimination “if it does not get in the way of connecting people in countries like ours”.

So in a country like ours, what does this really mean? Can and do these “mild forms” of packet discrimination co-exist in “packages where the users pay by byte – may be above a ceiling – and may not necessarily be proportionately”?

Sanjana

I did not want to get into this argument, because I have other things to complete right now but can’t help because there is a fundamental flaw in the argument you repeat.

You do NOT – I repeat, do NOT – pay more for your usage, as you claim – you pay only for the *promise* of a wider pipe – which you may or may not get. (That may be an issue for a consumer organisation which LIRNEasia is not.)

Speaking for myself, I never said you would get a better quality by paying for a ‘advertised’ wider pipe. In fact, research in UK suggests otherwise. Taken as percentages so called ‘high bandwidth users’ are the least likely to get what they are promised. (So be happy, you are not alone!)

[quote] Broadband customers are still experiencing connection speeds less than half those advertised, with the worst offenders being ‘high speed’ products, a study suggests.

On average, broadband speeds were just 48 per cent of those advertised, with the figure falling to as low as 26 per cent for high speed packages offering connections of 8Mb/s or more, according to the study.

By contrast, customers on 2Mb/s packages experienced average speeds of 1.8Mb – or 88 per cent of the advertised amount, according to broadband-expert.co.uk, a price comparison website.[unquote]

See more: http://lirneasia.net/2008/04/broadband-in-uk-speeds-less-than-half-those-advertised

What I always advocated was packages where the users pay by byte – may be above a ceiling – and may not necessarily be proportionately.

There is a difference between paying for bandwidth and paying for usage. No point continuing this discussion if you think both same.

If users only pay according to bandwidth operators have no incentive to deliver better QoSE. (That is is exactly your case) If users pay according to bandwidth AND usage (in terms of Bytes) they certainly have one.

That explains exactly why SLT Entrée outperforms SLT home in QoSE though both promise same bandwidths.

I’m curious then, what exactly would be the precise technical definition of your “mild form” of packet discrimination, esp. in the context of the regime as we have today in Sri Lanka, and how would it operate?

Is the proclivity towards “mild forms” of traffic discrimination a Lirneasia policy backed by research?

What do we do when this “mild form” of packet discrimination is exceeded and how to we, as consumers, measure it?

FYI, I didn’t say that non-discrimination of packets was a “paramount” policy objective, or one that “trumps everything else”, both of which were your formulations, not mine. I said it was as important as economics, costs and incentives in a country like Sri Lanka, not necessarily more or less so. I then pointed to the nature of the current regulatory climate and telcos as evidence that should the regime and telcos alone or combined decide on packet discrimination, there was very little to stop it from happening or ways of monitoring it when it does.

I then engaged with your assertion that taking broadband to the masses in the much the same way as mobiles to the BOP could be economically and socially most desirable, but dangerously does little or nothing to address the possibility of network management based on deep packet inspection and discrimination unless there were policies to check this.

On the economics too, I pointed to your own Ashoka Tissa I research that brought out the fact that high users paying more for internet access in Sri Lanka today actually have an inferior QoS than those who pay less. I asked you then what guarantee high users who are asked to pay more have of a qualitatively better internet experience without covert throttling of certain protocols and traffic?

Debating debating styles, to coin a phrase, must not gloss over vital questions that are unanswered by you and a conversation from which we all stand to gain.

Sanjana

I think you are talking to yourself and putting words in my mouth.

I have some urgent work to do. I will reengage when I get that done. In the meantime I would appreciate it if you rethink your debating style. I expected better.

I’m glad we are making progress. I completely agree with you – economics, costs and incentives matter. We also agree that as much as they do, so does the non-discrimination among packets. As a vital policy objective, net neutrality defines thus is important because I believe that packet discrimination for any user, whether she / he pays a premium rate, flat rate, subsidised rate or accesses the ‘net for free, is wrong.

I fear that regimes such as ours today could well use a populist policy measure of a providing broadband to the BOP to mask taking *their* version of the internet to the masses. Heck, they may even call it the Mahinda-jalaya. You may then have excellent access and throughput for all, but the nature of the “internet” that comes through the pipe may be questionable. Is this scenario really addressed by how much individuals pay for access?

Even the economics of it, if I understand you correctly, doesn’t quite make sense. You propose that high users need to pay more. Fine and acceptable in theory in Sri Lanka. But if I recall correctly, Ashoka Tissa I brought out that SLT’s ADSL “broadband” Entrée Package delivers better QoS overall than the Office Express Package, which I am on and pay a significant premium for ostensibly better levels of service. My example of personal experience with a leading Wimax provider actively throttling P2P traffic on their networks is another case in point of paying a premium only to face packet discrimination and deplorable QoS.

So we come full circle to my initial and fundamental query – what convinces you that users who can and are willing to pay a premium for high use actually will get any assurance of packet agnostic traffic management? Further, how assured can we be of those at the BOP are able to use the internet and the web, including all its protocols, products and services, in an unfettered manner?

As I note in my response to Deane in my blog post (http://ict4peace.wordpress.com/2008/09/21/net-neutrality-and-internet-qos-in-sri-lanka-redux/#comment-26121),

“In [the present] context, options available to ensure net neutrality range from public interest litigation, FR cases and groups acting as watch-dogs. For example, a Sri Lankan equivalent of EFF can help identify net neutrality issues, raise consumer awareness, hold telcos and their network management frameworks to public scrutiny (using tools like Switzerland) and use the media to name and shame those who don’t deliver what they promise.”

Sanjana

Net neutrality is a slogan. There is nothing sacred about it. As far as we are concerned, what are important are low prices for BOP consumers and incentives for ISPs to connect them. If our objectives can be achieved along with non-discrimination among packets (my working definition of NN), fine. But why would non-discrimination among packets (if that is your definition) be a policy objective that trumps everything else?

Economics matter. Costs matter. Incentives matter. We need to think these things through.

It’s not unreasonable at all. But I come back to my question. Are these pricing strategies for mobiles, translocated to broadband consumers, any guarantee of net neutrality?

You’re talking economics and I’m talking traffic management. Your propose that a model that services the BOP well today with mobiles will augur better broadband for all. I fear that even if that were true (and make no mistake, I hope it is) the resulting internet will, if we are not careful, be one that is tailored by telcos and / or governments to meet their parochial ends, be it profit or propaganda.

Sanjana

We’re talking business models and pricing strategies. Is it unreasonable to think that business models and pricing strategies that work in one set of telecom services have applicability in another?

When we talk about lessons for telecom from the selling of shampoo, we do not claim that there is anything common in the production of shampoo and telecom services.

Firstly, far as I know, net neutrality and packet discrimination don’t affect voice telephony on mobiles?

I fail to see the connection, much as you seek to draw it, between your emphasis on mobiles and the fact that net neutrality over broadband is already suspect in Sri Lanka. Voice telephony on the mobile surely different to packet discrimination on broadband? What in the growing usage of mobiles at the BOP gives you confidence in an economic model for broadband that guarantees net neutrality?

Getting people connected is great and welcomed. How we get them connected and to what are important – lest you are fine with the model of China and among other things, its singular approach to controlling internet access. As noted in my post,

“I guess given a choice between slow access to an unrestricted internet over blazing fast access to a restricted internet, I’d gladly choose the former. But is asking for both really that unfair?”

Thanks for the tip on Prahalad and as you know the 2002 article was co-authored with Stuart L. Hart. I wonder if you’ve read the more recent “The Next 4 Billion” from the World Resources Institute – http://www.wri.org/publication/the-next-4-billion?

Both articles strongly support, as you note, “extending [mobile] service to the bottom of the pyramid and in bringing prices to the lowest levels in the world”

Both don’t have a single mention of net neutrality.

Sanjana

Flat rate for all is a luxury only mature networks can afford. Better get people connected in some way than wait around for the unattainable ideal (which happens to be even more ideal for knowledge workers who write about these things).

Also just to note that, as you say, there are some assumptions and assertions made here by you and colleagues that don’t, in your opinion, need research to establish in fact.

That to me is a position. History may perennially prove you to be right, but that doesn’t make you objective or neutral. Why be defensive about it?

Sanjana

What is the well-spring of your optimism that a pricing model based on usage will improve the state of affairs in Sri Lanka for all consumers of “broadband”?

A concrete example – one of the offices I work at in Colombo has a Wimax solution for which we pay over 12,000 rupees a month. Skype and other P2P applications on it simply don’t work. Network throughput (up and down) is nowhere near what is advertised, at best. What is clearly a management framework that throttles P2P traffic suggests that (some) telcos in Sri Lanka today are getting away with charging premium rates for QoS they cannot and will not deliver.

I also believe in allowing operators and customers the freedom to figure out what kinds of pricing models work. But clearly, nothing in our market today works, as brought out by your own research to date. Speaking from the perspective of a minority of users who;

(a) can and will pay a premium to access the web at *advertised rates*
(b) are willing to live with a bandwidth cap plus dynamic throttling at peak times, a la Comcast in the US and slightly differently, Plusnet in the UK,

I’m yet to be convinced by your line of reasoning that, in Sri Lanka, pricing model based on usage will deliver better QoS for me, or address what is already a practice amongst some telcos in Sri Lanka to discriminate against certain types of internet traffic.

Sanjana

Take the second and third sentences. We have done the research on mobiles. We know what works. We do not know exactly what will work in broadband. But it does not seem such a controversial statement, based on what we have found. When I was Director General of Telecom, there were people who challenged me in public meetings arguing that the flat rate model of pricing found in the US should be imported here. I answered them more or less on these same terms. History shows us who was right.

What exactly do people mean when they throw around the phrase net neutrality? Is this about pricing models or traffic shaping? These questions need to be answered before we take a position.

But on allowing operators and customers the freedom to figure out what kinds of pricing models will work best, without compelling them to the rigidities of flat rate, all you can eat, we can take a position right now.

You say that “LIRNEasia is a research organization at core and that we do not jump to conclusions without investigating an issue fully.” This is a commendable. Yet in July ‘08, writing on Net Neutrality on this very site, you noted on that,

“If everyone is charged the same irrespective of use, what really happens is that the low-users end up subsidizing the high-users, especially in countries of the South, where the biggest cost driver is international bandwidth. What we need is a business model wherein low users pay only for what they use, in small amounts. This is what worked in mobile, and this is what will work in broadband. So let’s take the essence of the net neutrality debate, that no class of user should be excluded. Let’s not take the elements which argue that it’s wrong to charge for use/volume.”

http://lirneasia.net/2008/07/do-the-terms-of-the-us-net-neutrality-debate-have-relevance-to-low-bandwidth-countries/

The two statements seem at odds with each other. Please clarify?

Thanks,

Sanjana