NightKhaos on Digital Freedom

Digital Rights and Technology Blog

Archive for November 2010

There is a debate

leave a comment »

For those of you not following Question Time, please make sure you have a quick look at the #senateQT hash tag. You will notice a trend, the Opposition are going on about the NBN until the cows come home.

All because the Government did not realise a few documents regarding their NBN policy (hypocrisy anyone?). Honestly it’s really starting to annoy me, from both sides of the debate: what precisely is so important in the document that it is going to take you many months to redact it Labor? Just give them the document already, redacted or not.

And then there are the Coalition, who’s Broadband Policy isn’t, lecturing the likes of the Australian Senate almost continuously on how much of a “bad idea” this is, by any means necessary. And I’ve had it. Seriously.

Polarising the debate only serves to remove the best interests of Australians and instead create a situation of “he’s wrong, I’m right.” And why is this important? Because either way, someone in Australia will lose when it comes to Broadband.

Under the Coalitions Broadband Policy (the CBP), however, it my opinion that more Australia’s will “lose” than those who will benefit from it, and it is for this sole reason I support the NBN over the CBP. If the CBP were a measured approach to put forward the steps to ensure that all Australians receive an acceptable quality of service for what is more and more becoming an essential part of Australia life, such as:

  • incentives for Private Enterprise (PE) to build FTTH networks
  • mandates for PE to build open data, carrier and service neutral networks
  • mandates to prevent or reduce vertical integration within PE
  • the structural separation of Telstra and the generalisation of the minimum service mandate (in that a carrier must provide a service, not Telstra must provide a service to every home)
  • mandated restructure and complete review of Telstra Wholesale’s business in order to find and scape inefficiency as well as improve the quality of service
  • redefinition of what what these mandated services entail, such as voice capability either via VoIP with access to emergency services, or a POTS line, with a minimum bandwidth delivery or CDR to exit of network (not “Minimum Peak Speed”) of 1.5Mbps until by the of end of 2011, by the end 12Mbps at the end of 2015, 50Mbps by 2020, etc.
  • government investment in rolling out fibre and next generation wireless solutions to remote areas in order to meet the mandates specified by the above redefinition
  • government subsides for those who are unable to get fixed line broadband and are forced to use an expensive wireless service (such as Telstra NextG)
  • but most importantly, the acknowledgement that privatisation of Telstra has resulting in broadband service quality in Australia stagnating and it is a government imperative to invest in methods to accelerate the deployment of fast (50-100Mbps) services to all homes in Australia within the next decade.

Let me come back to a point I made earlier: “a minimum bandwidth delivery or CDR to exit of network”, many of you will probably already know what I mean, but for those of you who don’t I mean that the network is capable of providing, under normal usage conditions (it is impossible and prohibitively expensive to engineer a network that will always do this) provide a service of at least that speed to their Points of Interconnect (POIs) and within their network to every single user. Note this mandate should apply to all carriage service providers (CSPs) to ensure that there is little to no back haul congestion.

Now if the CBP was along those lines, rather than the patchwork policy they offer now, I would support it over the NBN considering it’d likely cost the government less than “going it on their own” as Labor are doing. But it’s not, not even close, if anything it seems to say all the right words, but in reality things like “minimum peak speed” and investment in HFC over FTTH mean that it will leave a lot of Australians with slow, unreliable broadband. In fact, the CBP isn’t actually that well thrashed by the press so no one can be exactly sure what the CBP will do, because it is much easier for the likes of Mr Malcolm Turnbull to say “look at Labor, they are wasting money” than to say “look at our policy, we are doing nothing to fix the problem” isn’t it?

What about the NBN? Well as I pointed out in my previous post the NBN has some flaws to it’s implementation that mean people like @philhart will be forced off a perfectly reliable, albeit it slightly slower than I’m sure he would like, ADSL connection, and put onto a Satellite connection because the NBN is going to completely replace Telstra’s Copper Access Network (CAN) without even considering, for a moment, that there may be some circumstances where leaving the CAN in place would be better for the consumer. Granted the number of users in his position are scarce, but I thought the point of the NBN was to give better broadband for all Australians?

All Australians deserve access to good quality, cheap, broadband services, not just the majority, under the NBN, and not just the minority, under the CBP. I don’t like picking better the lesser of two evils, so if you really want to get my attention Coalition, put an actual policy on the table.


Written by NightKhaos

November 23, 2010 at 10:11 am

Do we really need FTTH right now?

with one comment

It’s a question that has been bothering me, not because I question the need, in the long term, for FTTH (Fibre to the Home), but because I’m thinking about the short term as well. It comes down to this: does any solution we implement to fix broadband in Australia other than FTTH have a viable upgrade path to fibre to the home that justifies spending less in the short term even through the in the longer term it will likely cost us more?

Another question that has been bothering me is “does the NBN provide a good mix of FTTH and other solutions for the whole country?” since these two issues are linked; is the NBN just another interim solution wrapped in a $43billion price tag, or is the majority of money we spend on the NBN going to result in savings in the long term?

You’ll notice, rather early on, that I have separated the issue of FTTH and NBN. I feel it is important to reiterate this point because on Twitter there seems to be an assumption that the two things are one in the same; they are not. The NBN is just one way to deliever FTTH to majority of Australia’s and I’m increasing getting the feeling that it is the wrong way.

Of course, considering the alternative (or lack there of) offered by the Liberals, spearheaded by the ever increasing short-sightedness of Mr Malcolm Turnbull, if we want, and believe me we should want, FTTH, we need to do it via the NBN.

To start this consolation of ideas and frustrations I’ll begin with the argument I hear so much of in the Australian Broadband Debate, speed. Speed seems to be all it has come down to, on the one hand we have the Labors offering so much speed we don’t know what to do with, 1Gbps to each home connected to their FTTH solution, with 12Mbps wireless to the country. I can see a couple of my fellow geeks drooling at the prospect, and to be honest, I don’t really blame them. Us geeks are the first to know exactly what we are going to do with that bandwidth.

Then there is the Liberals, taking a more conservative approach, which makes sense, being a conservative party, offering what seems like a much more reasonable goal of 12Mbps to everyone. Which sounds fine and dandy, and also conveniently is pretty much, with a few tweaks here and there, what we can get now.

And truth be told, I’m sure the majority of people have no idea what 12Mbps will give them. Because even in the days of dial-up there has been a complete disconnect between the speed, and how the technology can be utilised, and what that means for consumers. The dreaded “Up too.”

Of course there has been attempts to make it easier for consumers to understand what that means but all things considered, they still have no idea. And that is just fine to be honest. However a consistent delivery is something I would like to see. What do I mean by that? Well, remember this when you first got 256k ADSL? “Over four times faster than dial-up” I’m sure you were told, and to your surprise, it most definitely was. You couldn’t believe the difference. Then you got full speed ADSL, and the introduction of the “up too” came back, just when you thought you’d got rid of it with broadband it comes back to haunt you.

This is where FTTH will shine. If I have a 25Mbps plan, and I upgrade to 50Mbps, I can most definitely expect the websites to load around twice as fast as before. And this is by far the greatest thing that could happen to the telecommunications industry. Let me put it this way:

If you currently have a connection capable of streaming in real time a single 720p video from a VoD service, like YouTube, and you are currently on the 25Mbps plan, you would reasonably expect that if you upgraded to 50Mbps, you could quite happily stream two.

And to extend to this, if your friend across town can stream that video on his 25Mbps connection, wouldn’t you reasonably expect when you get home that you could too since you too are on a 25Mbps connection?

It’s this sort of assurance that we can’t get with ADSL and wireless. It’s fine now, that most people only go online to check Facebook. But like every new technology, like smartphones, once someone makes it easy to use everyone will catch on. I’m sure one day we’ll get a VoD service in Australia that is what the iPhone is to smartphones.

Kinda of like Hulu and Netflix have started to be in America. Of course, a few in the industry, like @jimboot, have stated that compression make this idea redundant; and in a lot of ways he is right, we don’t need all that much bandwidth to achieve it. A high definition video stream in 720p of reasonable quality using only 4Mbps.

The only problem with this is that if everyone in the family want to watch a different stream, it starts to add up, and we only have enough bandwidth for three streams per household under the Liberals policy. Which isn’t that bad, but compression can only do so much. As 1080p and 3D content become more popular as the technologies mature, we will need ever more bandwidth, and that 12Mbps might no longer be enough for three streams, in the worst case it might not even be enough for one.

Then the the agruement for upload. Currently the majority of technologies deployed are asymmetric, i.e. the amount of download far exceeds the amount of upload because the majority of home internet users are consumers, not producers. This is unfortunately changing, which means network engineers need to rethink the services they are delievering.

Let us take Facebook, Flickr, Picasa, YouTube and various other image and video hosting services. They make it very easy for users to share content. So much so that users are uploading far more content than they were before. The majority of the content for home users is of the downstream type, but the trend is changing.

It was common through when the technologies like ADSL and HFC cable were developed that users would want to download far more than they upload, and this has proved true with a few exceptions. Those exceptions include video conferencing, working from home, and online gaming. All activities that are getting increasingly more popular.

So this assumption lead to the development of the 20 to 1 rule of thumb. In order to download at 20Mbps you will need to provide 1Mbps of management traffic on the upstream. The problem with this is obvious, the rule has very little margin for error. So if your son upstairs is exhausting the connection downloading a 720p video from YouTube, you are unlikely to be able to do even low traffic operations like a VoIP call via Skype.

Now many of you may have guessed that I am going to point out to you with Fibre we will get an increase to the upload bandwidth. Of course there are two fundamental issues with this: of the plans I have seen so far from the likes of Internode and iiNet they are still not offering products with higher upload ratios or symmetrical (the upload rate is equal to the download rate) on the NBN, and for services (with the exception of Wireless) all can easily be adjusted to provide greater upload or even symmetrical service (at the cost of reduced download).

This is where the NBN shines out like a torch in the night: the costs of changing the ratio are far less for the NBN than they are for existing ADSL technology. The reason for this is that the NBN network is being implemented as a dumb pipe system. You can tell it what upload and what download you want for the particular service and it delivers it. You can even install more than one service per pipe, say if you wanted cable TV and 100Mbps/25Mbps internet for example. The NBN can quite happily provide this configuration.

However with ADSL2+ technology we are already stretching the limits of the technology. The best that can be down is a 2Mbps upload (only offered by Internode at the moment) and to adjust the ratio further than that, not only would the ISPs need to replace the very expensive DSLAMs at exchanges (each DSLAM costs in the order of $75K) but also get the customers to change their modems as well ($100 per customer).

HFC is a little bit more flexible, fortunately, but unfortunately cable providers are very hesitant to adjust the ratios because it would mean they could no longer offer customers the brand new “100Mbps” plans via DOCSIS 3.0. Because, like I stated above with downstream, there is even more of a disconnect to what upstream means. With FTTH since the systems are cable of delivering a total bandwidth of around 1.0Gbps between the splitter and ONT (or the last few hundred meters as it were) the providers have much more options. They can even provide a 100Mbps/100Mbps service without any worry.

Recently, even more interestingly, have been the development of “cost saving technologies” since the industry seems unsure on Fibre (the industry has been unsure of Fibre for about a decade if we’re honest, I read an article recently, written sometime between 2002 and 2005, that proved that, especially for new offices, the cost of FTTD (Fibre to the Desk) is actually only about $9 more expensive per desk than running a traditional Cat5 network even through it involved installing $200 fibre cards in each desktop.)

You can check out that article here. In the same article it did point out that FTTH solutions are approximately 10x more expensive than copper based solutions, and this seems oddly true even today, however it did make an interesting point about remote areas where it will be cheaper to run fibre because of the greater performance characteristics. I find it slightly ironic that the NBN plan opts for wireless for remote areas in light of this, but I never said the the NBN was a perfect plan.

A recent article suggested a new technology that could be used as the Fibre killer could be in the form of VDSL loops, able to provide approximately 25Mbps total throughput (both upstream and downstream) to each home in a loop of 16 homes with a maximum of 400Mbps per loop. And it could be a possible viable interim solution, as to be done properly it would require that fibre drops be run out to each cluster of 16 homes. You can read that article here.

However my friend pointed out to me that ring structures in networks tend to have very high latancy, so the online gaming community may not be so appreciative of this “upgrade.” I also have a strong feeling that such a “middle step” is a waste of time, but that won’t stop the Liberals from suggesting it, and good on them, on face value it actually looks as if it might work. If I wasn’t so familiar with the technology behind it I might have gone for it too.

When it comes to the NBN through, I am slightly concerned, another person I follow on Twitter, @philhart reports that althrough he is currently on a reasonable ADSL connection, under the NBN he will be downgraded to Sat. (And yes, even through he may get 12Mbps once the two new Sats have been launched for NBNCo, the latency performance of Sat based internet connections is such that I do consider it to be a downgrade compared to a 1.5Mbps internet connection).

So what does that mean? Well it seems to me that we do need FTTH, but maybe not exactly the FTTH that the NBN will provide. But it’s a new policy, it’s bound to have a few teething problems. Let’s hope people like @philhart can also be taken care of in Broadband Policy, for example, why don’t we keep part of the CAN (Copper Access Network) functional where to replace it would not be cost effective in the medium term rather than just saying “No, you’re getting Sat”?

It’s suggestions like these I want to see coming out of the mouths of people like Mr Malcolm Turnbull, not assertions that they can do it better without FFTH, because I don’t think they can.

Written by NightKhaos

November 11, 2010 at 8:39 am

Posted in NBN, Technology

Tagged with , , , , ,