http://www.abc.net.au/news/s
Seems they've changed their tune since this announcement
http://www.inter
Given that this is a phase 1 trial rollout of the NBN it's to be expected there'll be a few teething issues.
That's the whole point of a trial rollout, to iron out all the kinks before it's released everywhere.
I'd pretty much assume that Internode's network engineers are working very closely with the NBN team to fix any problems that might arise.
Or, the ABC is just being biased and creating issues where there are none. From what I can tell, their reporting on the NBN hasn't exactly been neutral.
"My technician tells me it averages about 30 megabits per second instead of the 100. So while things are faster, it's not significantly faster."
$20 says his 'technician' used speedtest.net lol.
Or, the ABC is just being biased . . .
My goddness !
The ABC biased ?
Who ever would have thought this !!!
More news here, not on the ABC:
The Australian has been a staunch opponent to the NBN. Their reporting has definitely been biased throughout the entire debate.
It wouldn't surprise me that the cable under the Bass Strait has become congested already. With all of these premises being hooked up to high speeds it was only a matter of time before it needed a capacity upgrade.
It wouldn't surprise me that the cable under the Bass Strait has become congested already. With all of these premises being hooked up to high speeds it was only a matter of time before it needed a capacity upgrade.
With due respect to Internode, you'd think that the problem would be self-evident and they'd be onto the issue and be purchasing more capacity.
Unless there is no more capacity to purchase?
I'd be certain that Internode have it's hands tied by a bottleneck outside of its control. I'm sure they wouldn't be deliberately holding the speed back, it doesn't make sense from a PR perspective...
ABC is normally accused of bias by the liberals who are antiNBN.
The one thing the NBN will expose is how ill prepared the current Australian backhaul and international links are. Some are within nodes control but what server was the speedtest done on?
I would presume peering ISPs are redesigning their interchanges in preparation for the NBN.
So where is the bias in The Australian's story?
Not talking about this one in particular, although it is a negative story about it, talking about their overall coverage of the NBN. Just take a look back at a few of their stories and it becomes self evident.
The one thing the NBN will expose is how ill prepared the current Australian backhaul and international links are. Some are within nodes control but what server was the speedtest done on?
This has been done to death in the NBN forum, there is enough capacity for everyone. The Telstra Endeavor has barely any of it's capacity lit, same is said for PPC-1, Southern Cross Cable etc. There is enough capacity going to the US, and there is a new link supposedly going up ~2013 anyway which will have a capacity around ~4 terabits. It's been a while since I've looked in the NBN forum but that's what I remember. The only thing that is lacking is the SMW3. we need another cable like it going from Perth to London.
Anyway I wouldn't be surprised if the "technician" was using 10Mbit ethernet, or if the kids were the bottlenecking the network with their own downloads.
there is enough capacity for everyone
I think the poster was (also) referring to the backhaul from here in sunny Tassie up to the North Island. We only have two cables linking us to them and one is owned by Telstra. :(
~ Edited to make more sense.
I think the poster was (also) referring to the backhaul from here in sunny Tassie up to the North Island. We only have two cables linking us to them and one is owned by Telstra. :(
yes, but if there was congestion on Nodes backhaul from Tas to the mainland we'd all be suffering from it down here, no such issues as far as I can see.
The issue is more than likely in the TNBN network before it gets handed off to internode, or there's something local at the school in question.
With due respect to Internode, you'd think that the problem would be self-evident and they'd be onto the issue and be purchasing more capacity.
How much money should they pour into this NBN test?
100mbits of IP into Tasmania – I have heard costs in excess of $150 per mbit.
100Mbit x $150 = $15k per month
There will obviously be a level of aggregation to alleviate most congestion issues, but the problem is the average speed difference between NBN customers and DSL customers mean that capacity constraints seem harsher to NBN customers.
Until a greater level of scale is achieved with more 50 and 100mbit customers connected, I suspect most people will only occasionally peak at their max speeds.
Unless there is no more capacity to purchase?
Basslink was only connected pretty recently, and there is still the Telstra cables – I would think it is very, very unlikely that both links are full.
Jason
Given that it's a school, and a religious school no less, it wouldn't surprise me if they were running internet filtering equipment that was unable to keep up with the 100Mbps connection.
So where is the bias in The Australian's story?
I would be wary of a quote by John Harris (a PR person not directly employed by Internode) rather than by someone who actually works for Internode and knows what they're talking about.
looking at nodes network setup seems like pipe have a hand in it mmmmmwith all the bull put up by them lately u have to wonder on this if it was tpg then mmmmmmm and how many are complaining of sub 100 k speeds on tpg right now a lot
network goes to hobart then conects to pipe from there
not saying they are doing it but makes u wonder a bit
So where is the bias in The Australian's story?
In the title National Broadband Network can't keep pace at Tasmanian school .
According to its report (if true), it is not the NBN infrastructure that cannot keep pace, it is Basslink.
It is like missing a flight because your car broke down on the way to the airport and then complaining that you did not get to your destination on time because Quantas service was unreliable.
The issue seen by our customer in this case is not something obvious, and its not actually a link capacity issue on our side of the fence, to any extent that we've yet discovered (and we've been looking for one for a while now in case there is one).
i.e. no, Basslink is not 'full' in a general sense, and during school hours we have many hundreds of megabits per second of available capacity on our Basslink paths to Hobart from Melbourne.
The school link works fine in practical terms, but its not showing the expected peak download speeds for single threaded downloads. Way above ADSL2+ rates, but not magically '100M on demand' right now.
Some of the claims in the media seem to be a bit over-simplified, in looking to declare some specific 'smoking gun' before we've actually had a chance to locate where that smoking gun might actually be.
There's nothing intrinsically wrong with the technology and there's no underlying resource starvation in our backhaul links that we can see... thats the point.
One thing people will have to get used to with 100M links is that a chain *is* as strong as its weakest link, and at those speeds, that weakest link can be something invisible at lower speeds (like duplex mismatches on ethernet ports, slow throughput to disk on client side machines, overloaded source servers on speed testing sites, TCP window sizes not being tuned properly for these very high speeds ... all sorts of things matter much more at 100M than they do at 10M).
It only takes one link in that chain to have even a tiny level of packet loss before a single threaded transfer takes a bit hit in throughput (but typically with no impact on other transfers also happening at the same time).
The source of that, as above, can be one of many possible things – of which the above are only a small subset.
One of the main benefits of 100M links is in fact that sort of effective concurrency – lots of people accessing content at once without getting in each others way. And that seems to be happening entirely as expected for our customer.
Its not all that uncommon to see single threaded throughput limits in high speed, long distance paths that still allow lots of machines to hit very decent speeds in parallel that aggregate up to the full link capacity with ease.
Comments attributed to the school notwithstanding, our staff have actually been in continued routine contact with IT staff at the school for some days now (and were working with the school on resolving their concerns before and during the media that appeared today).
We're happy to get to the bottom of the issue (which we're told is having no operational impact on the operations of the school in general – the link is comfortably exceeding ADSL2+ capabilities in the meantime, as you'd expect).
But we'll have staff on site next week to sort it out completely, one of whom is setting up to work along each stage of the data path from Hobart to the school, working with NBN technical staff, to figure out where the gremlin is (if there is indeed one in there rather than it being an end system or router issue).
We are as interested in the NBN working brilliantly for our customers as they are. So its not an exercise in apportioning blame.
Everyone is on a learning curve with the NBN, and this is simply a part of that learning curve. NBNTas is a very new network and every link in that chain has barely got the paint dry on it (end to end) – and there are many links in that chain.
Meantime: No emergency. Our customer is seeking to understand why the link isn't hitting the single threaded download rates they expect, and we're just as interested to work that out as they are – so we will do so.
We are very committed to high performance services. We always have been. The NBN takes that to a new level for everyone. And we have a great network to show the NBN to its full advantage.
We all just need to remember that the paint is scarcely dry on NBNTas as it stands – it was built very quickly, and we're all learning (heaps, in a good way) from the exercise – and from our continued involvement with it.
I'm sure both we and NBNTasCo will learn something useful from the exercise of getting to the bottom of the issues at play in this specific case – and we'll use it as a source of useful learning to help our current and future customers get the most of out of their high performance Internode NBN services.
Cheers,
Simon
Thanks for the detailed reply Simon
give us laymen a little incite on how the NBN can have probs
Everyone is on a learning curve with the NBN
Isn't that's why its a trial
Doesn't the media understand that
In the last line of the article the NBN spokeslady said it probably had to do with cabling which was the ISP's problem.
Is she just passing the buck because no one is allowed to detract and deride the precious NBN?
:D
Not when there is a story to be had. Something the industry had better do is make sure actual throughput expectations are properly communicated. The media will have a feast if they do not and the liberals will destroy the project on coming to power. Sadly the communications industry is awful at setting realistic expectations and giving good service. Node is the cream at the top of some very sour milk.
Doesn't the media understand that
Biggest question for me is why the school felt they needed to make big deal out of it in the media, rather than continuing to work with Internode to resolve the problem.
Biggest question for me is why the school felt they needed to make big deal out of it in the media, rather than continuing to work with Internode to resolve the problem.
It would not surprise me if someone from either The Australian or the Liberal party made contact with as many known early adopters (excluding those who repeatedly praise it such as Bob of Midway Point) as they could trying to dig up some dirt.
Biggest question for me is why the school felt they needed to make big deal out of it in the media,
I'm going to guess here... but it's plausible that the media approached the school first, and asked them how it was going (the NBN thang being the latest rage and all that, down there in Tasmania.) Some poor, frustrated IT sod working at the school said it wasn't as good as it ought to be, and the media ran off with a story.
But we'll have staff on site next week to sort it out completely, one of whom is setting up to work along each stage of the data path from Hobart to the school, working with NBN technical staff, to figure out where the gremlin is
As I was reading Simon's long post I was thinking that if only testing could be done at each point along the way to narrow down what/where the problem is and there it was in black and white.
This is the reason people join Node.
im going to bet that the servers they connect to (most on the internet) have a shit configuration of the TCP/IP stack and is ACKing every 2nd packet add in the latency for that ACK and watch single TCP session performace die in the arse.
I see it a fair bit on Long Fat pipes.
I'm sure both we and NBNTasCo will learn something useful from the exercise of getting to the bottom of the issues at play in this specific case – and we'll use it as a source of useful learning to help our current and future customers get the most of out of their high performance Internode NBN services.
Thanks Simon:
Is there any possibility you could add an item on the resolved workflow for this job to post an update for us as to what the cause was, I'm sure a lot of us would be interested, and its a sure bet that the newspaper wont carry it.
Perhaps this very subject could even be wrapped into your ABC radio chat somehow.
What I find interesting about this is that out of all the customers signed up, only one (so far) what appears to be speed issues which to me seems a pretty good success rate.
Still something that needs to be fixed of course, but still not a bad success ratio overall.
i.e. no, Basslink is not 'full' in a general sense, and during school hours we have many hundreds of megabits per second of available capacity on our Basslink paths to Hobart from Melbourne.
Internode's John Lindsay said customers' speed problems were caused by capacity constraints, which the company was "working hard to solve".
The ISP had bought more backhaul capacity to ease the congestion, he said. "Aurora Energy is providing us with a capacity upgrade today to improve our network performance."
So...Simon has posted that there is no capacity problems, while obviously not checking, because John has been speaking to the media telling them that they dont have enough backhaul – and that is the cause of the customers speed problems.
This one is difficult. Simon ?
Simon has posted that there is no capacity problems, while obviously not checking, because John has been speaking to the media telling them that they dont have enough backhaul
There is more to Capacity than just backhaul, when we are talking about 100mbit/sec. The capacity of the sorce server to provide the stream at that speed, TCP RWIN tuning etc..
There is more to Capacity than just backhaul, when we are talking about 100mbit/sec. The capacity of the sorce server to provide the stream at that speed, TCP RWIN tuning etc..
Why are you trying to spin it? Dont defend it – John has already admited the problem.
The issue is simple – Basslink charge far too much for access to their cable. If they reduced the price to something like 1/5th what they charge now then you might see capacity constraints ease to something more normal.
The school link works fine in practical terms, but its not showing the expected peak download speeds for single threaded downloads.
...
One of the main benefits of 100M links is in fact that sort of effective concurrency – lots of people accessing content at once without getting in each others way. And that seems to be happening entirely as expected for our customer.
So let me get this straight. The problem is that a single thread download is, for some as yet unknown reason, limited to 30Mb/s. Multi-thread, multi-application and multi-user usage can easily saturate the link with as little as four threads. That number can be generated easily in most household usage pattern, let alone the usage pattern of a multi-user environment like a school.
The supplier has been checking into the issue and has committed to sending staff on site to diagnose and fix the issue, which really is not a big deal in that environment.
All I can say is – What a media beat up! This is not even a mole hill, maybe an ant nest.
Oh, and Simon, you should ask your PR firm to not comment on anything technical. He makes it sounds like the problem was caused by Basslink congestion, which could sound like blame shifting if that is found to be untrue.
Why are you trying to spin it? Dont defend it – John has already admited the problem.
Not spinning, just stating a simple fact, or do you disagree and say that only backhaul congestion could be the only reason this problem is occuring?
The issue is simple – Basslink charge far too much for access to their cable. If they reduced the price to something like 1/5th what they charge now then you might see capacity constraints ease to something more normal.
That may be true, perhaps you have access to charts and such that show the traffic being carried by this link and it's overloading? Perhaps you would share this information with us.
Or perhaps your guessing??
I have no reason to spin this, I'm not a Node rep, and I don't live in Tasmania.
The issue is simple – Basslink charge far too much for access to their cable.
Simon said it isn't basslink- and non NBN based connections not having issues would bear that out. The Aurora link being upgraded will be the interconnect or backhaul within the NBNTas network.
Thats not spin & its still nothing to do with basslink.
Simon has posted that there is no capacity problems, while obviously not checking, because John has been speaking to the media telling them that they dont have enough backhaul
Simon was talking about Basslink (Tasmania-Mainland) and I'm assuming by talking about Aurora Energy that John is referring to backhaul from the NBN's TAS POI to Hobart (but I could be absolutely wrong — perhaps its from the trial site(s) to the NBN's POI...)
(my views do not represent the views of the company, and I in no way know anything about this situation — lovely to have a bit of rain this morning though)
More news here, not on the ABC:
http://www.theaustral
Internode gets a mention.
The usual "It wasn't me Sir, it's him!" finger pointing. It's a wonder Tel$tra didn't get a mention and get accused of hacking into the Bass Strait underseas cable and siphoning off the other 70mbps :)
I suspect we will have a lot more of this uninformed nonsense before the "trial" finishes. None of the media reports appear to accept that it is a trial and that teething problems are part and parcel of that process. What do people expect – a 100% perfect and fault-free trial service on day 1? Even when the trial period ends and the NBN becomes an officially launched roll-out, it won't be a perfect turn-key operation.
It's just more fuel to be added to the fire while that other mostly uniformed political debate rages on elswhere in Canberra.
Meanwhile back to my experimental NBN rollout: 2 jam tins connected with heavy gauge fishing line and an Eveready torch on each end flashing Morse code :p
by talking about Aurora Energy that John is referring to backhaul from the NBN's TAS POI to Hobart
That would probably be the transit from the NBN PoI in Cambridge back to Internode's gear at the Pipe DC in Hobart?
And for the record, John can step up and claim his honorary Tasmanian membership card at any time – blaming Aurora is what any real Taswegian would do in this case! :P
Another story again:
"Internode's John Lindsay said customers' speed problems were caused by capacity constraints, which the company was "working hard to solve"."
That would probably be the transit from the NBN PoI in Cambridge back to Internode's gear at the Pipe DC in Hobart?
Pipe building houses HBA1, but where's HBA2? :-P
Another story again:
"Internode's John Lindsay said customers' speed problems were caused by capacity constraints, which the company was "working hard to solve".
of all the problems to have, this is a good one. It shows not only that people are prepared to take on the 100meg connection, but that Tassie needs it's backhaul upgraded as well.
As for the connections to the US, those of us who've been on the internet for more than 5mins will understand backhaul issues back to the US. But far be it from The
Australian to try to educate it's readers.
So...Simon has posted that there is no capacity problems, while obviously not checking, because John has been speaking to the media telling them that they dont have enough backhaul – and that is the cause of the customers speed problems.
This one is difficult. Simon?
Its only difficult if you are starting from the premise of my being either uninformed or a liar.
I'm neither.
Its also a bit illogical, given that other customers on the NBN network concerned don't have the same performance issues, despite being connected via the same interconnect link into NBN Tasmania in Hobart. Downstream of that point all data flows are delivered by the NBN network itself (beyond any link that Internode provides or controls).
i.e. There are a number of links in the data chain concerned in the initial NBN deployment in Tasmania. As there are with all networks.
One link in the path that Internode does control, and which is provided by a third party supplier, was upgraded on Friday at my request, and it was John who arranged that upgrade.
Hence, far from my 'not checking', Jason: I asked for it to be done.
I asked for the link to be upgraded to simply eliminate that specific link as a potential suspect.
Because the media were looking for a story, and a simple smoking gun, I guess they took John noting (in response to an enquiry) that we'd done a link upgrade, as being the necessary smoking gun they wanted to find.
Except that it probably isn't.
Basslink, another link in the path concerned, has hundreds of megabits per second of headroom during school hours. So thats not the smoking gun either.
Other customers on NBNTas that we work with routinely aren't seeing the same performance issues that have been asserted to exist at the school concerned, meaning the problem, if it exists in the network, logically exists in the NBN Tasmania part of the network or in the client site itself (as I have indicated earlier in this thread).
Or it could be something else entirely, such as a very low incidents packet loss fault somewhere else in the chain (even a tiny level of packet loss can throw a high capacity single threaded TCP transfer back into the 'slow start' TCP realm and need it to ramp back up again over a few seconds. Meantime other concurrent data transfers are unaffected).
So the only thing that is 'difficult' here is to understand why there is such a degree of interest in such a routine thing – that being a customer with a reported link performance issue that Internode is in the process of working through with them in the normal way.
Here's something to appreciate:
If the link that was upgraded on Friday was the problem, we wouldn't still be sending a member of our staff out to work through the entire end to end data chain from Hobart to the school concerned during this coming week.
We still are doing that (thanks Brad).
If the link that was upgraded on Friday was the problem, we also wouldn't have other customers able to hit essentially full line rate on their services on the same NBN Tas network at the same time.
I'm a little mystified as to why a single customer with (at this time) an undiagnosed performance issue is such big news. As I've said earlier, there are all sorts of reasons why peak single threaded performance can be impacted on high speed links, and a great deal of those reasons can mis-represent as being a 'network' fault when in fact they're client or server faults.
We love providing excellent services for our customers and we are both keen and happy to investigate in order to fully understand the nature of the specific problem here, in order to both resolve it, and also to learn more about the sources of both perceived and actual performance issues in the NBN context, so we can apply that learning to future customers and future network expansion.
The NBN is very new. We are all learning – and enjoying the learning experience because we like learning new things. Thats another reason we're sending someone out to work right down the chain next week in person – because we love to learn things, as it helps us provide better services in the future. This sort of domain knowledge is invaluable as the NBN gets built and expanded in the coming years.
NBN technical staff in Tasmania are assisting us with the work to be done in this coming week because they are, of course, just as interested as we are in ensuring that all access experiences on this network are great ones.
There is simply no conspiracy here. We're dealing with a customer who wants to either see their service work better, or to understand what is missing in their perception of why they think it isn't good enough now.
Thats great, and we'll look forward to the results of investigating this over the course of this coming week, which will yield the results they (and we) seek in that respect.
But the bottom line in regard to the link upgrade concerned, to come back to that is:
- I asked John to arrange and expedite that upgrade on Friday with the relevant supplier and our network operations team (and this was done).
I asked him to do that in order to eliminate a theoretical source of issues (not withstanding that our network monitoring didn't, and doesn't, show that link to be overloaded before or after that upgrade).
- John informed the media that we'd done a link upgrade because when you are being called by people looking for a bone, you throw them a bone. He accurately stated something we'd done on Friday in response to demands to say what we were doing about 'it'.
But 'it' isn't actually that simple.
None of that makes me a liar, in identifying a sample of the other possible causes of the customers perceived issues, and it doesn't make a single site network fault into a network-wide conspiracy either.
Its also worth appreciating that we aren't fans of finding someone else to blame as the instant response to media queries about our network.
Its entirely possible the true underlying issue here exists in the client site, in the NBN Tas links between Hobart and Smithton (which are a long way from each other), in a network interface port somewhere, or in one of several other spots.
If its actually a capacity issue in any Internode controlled link, then its a capacity issue that isn't showing up on any of our capacity graphs! Nonetheless, if it turns out (somehow) to be something in a link we operate, we'll fix it- just as soon as we find it.
We aren't sitting on a smoking gun in the meantime.
Its both inefficient and (at a deep level) a bit silly to expect us (or anyone else) to declare with certainty 'who to blame' when the underlying issue has not actually been diagnosed as yet.
At some level the issue has not even been clearly characterised as yet (the link isn't down, and a speedtest.net result below expectations doesn't actually provide a link fault exists either – meantime the school is continuing to routinely use their network link to get routine work done – just fine).
Maybe its been a slow news week. Or maybe there is an agenda for some in the media to try to frame the NBN itself as somehow fatally flawed in technical terms (hint: it isn't).
I'll point out again that our NBN services in general are working just great, and that locating and resolving network faults are a routine aspect of running a network.
No conspiracy. Merely some strong interest in making our customer happy (its a strong motivation for us – always has been), and in understanding what is going on at the school in performance terms.
To that end, again, we are sending a member of our staff on a mission this week to find that out in person. He's starting from Hobart (where we have customers on the NBN running without the same fault) and working his way over to Smithton, testing along the way. He'll then go over the entire configuration and hardware on the customer site.
We've not been interested in participating in a breathless blow-by-blow discussion with the media of exactly what our staff are doing to diagnose a customer specific fault. Thats just not productive. And I think its actually a bit silly.
We have a preference for fixing things first and then indicating what the problem was. Doing it the other way around isn't very logical.
Regards,
Simon
I suspect we will have a lot more of this uninformed nonsense before the "trial" finishes. None of the media reports appear to accept that it is a trial and that teething problems are part and parcel of that process.
+1
With a side order of there being a learning curve here – for customers as well as network operators – on what to expect from a 100M link in real world terms.
We're enjoying the process of being involved as a key and day one provider of services into the NBN.
One of the outcomes of this 'stage 1' network is going to be some work on properly defining, for our customers, what to reasonably expect from their services – and the impact on the performance of those services that sheets home to factors that are quite invisible at lower speeds but which can be extremely significant at these data rates.
Its very easy to have even minor factors impact the single threaded transfer rate over a high speed link.
But one of the nicest things about 100M links is that you can run lots of transfers of data at once without them tending to impact each other, as they would on lower speed links. Ideal for (for instance) a school...
One thing people will have to get used to with 100M links is that a chain *is* as strong as its weakest link, and at those speeds, that weakest link can be something invisible at lower speeds (like duplex mismatches on ethernet ports, slow throughput to disk on client side machines, overloaded source servers on speed testing sites, TCP window sizes not being tuned properly for these very high speeds ... all sorts of things matter much more at 100M than they do at 10M).
Which meshes nicely with my post from the closed thread filled with people decrying the skills of the school technician or coming up with excuses ad infinitum why there is no problem and everyone is just too stupid to get it working right...
eg.
****************
Obey The Fist! writes...
As far as we know the "IT guy" working for the primary school (he must have flunked every class in the industry to score working for a primary school as an IT job) is trying to download from limewire with only one peer or something.
LoM writes...
He's technician tells him it averages 30Mbps but if he's like most "technicians" (ie N00bs) that I know, he will have no idea how to conduct a proper throughput test.
Smells like a beatup alright.
I guess you guys didn't read down to:
Mr Bakes says the internet service provider has been trying for months to fix it.
So yeah, there's no problem but the ISP has spent months trying to fix it... why? I guess the ISP's technicians must have flunked every class in the industry as well...
WP posters, giving The Australian a run for it's money in 'how low can you go'.
KingForce writes...
I don't understand why the principal expected 100 Mbps to access sites outside of Tasmania.
Because Conjob and the menagerie of adherents have been preaching the good word of the NBN in such a hamfisted fashion that people who don't know better (ie. most of them) seem to be under the impression that as soon as you get 100 Mbit, you can access 100 Mbit full time to anywhere...
****************
Anyone who works with networking knows that the slowest choke point determines your overall bandwidth. 100 Mbit has been spruiked up to the point of mythic proportions but you'll rarely achieve it on any one download or process. You can certainly have more stuff running concurrently but even that is limited if the choke point is between your ISP and you (or with your local equipment etc).
This has become a story purely because the expectation has been set so high to begin with. Real world performance (kinda like those NextG theoretical speeds vs actual performance) will almost always be lower than the max number available, even on 'all singing all dancing' fibre optics.
KingForce writes...
Because Conjob and the menagerie of adherents have been preaching the good word of the NBN in such a hamfisted fashion that people who don't know better (ie. most of them) seem to be under the impression that as soon as you get 100 Mbit, you can access 100 Mbit full time to anywhere...
****************
Anyone who works with networking knows that the slowest choke point determines your overall bandwidth. 100 Mbit has been spruiked up to the point of mythic proportions but you'll rarely achieve it on any one download or process.
Maybe someone should tell Conroy this...but of course he knows full well, but conveniently forgets when he says things like:
"well unlike other technologies (wireless/adsl), when you buy X Mbs on the NBN you get X Mbs guaranteed!"
Have seen him use that line a few times now.
"well unlike other technologies (wireless/adsl), when you buy X Mbs on the NBN you get X Mbs guaranteed!"
And that is accurate as he would be talking about connection or sync speed not file transfer speed.
Both wireless and ADSL are best effort so your actually sync speed can vary wildly depending on many factors. Fibre should always "sync" at the speed you are paying for unless there is a fault. The problem is most people assume that sync speed = download speed where this is not actually the case.
And that is accurate as he would be talking about connection or sync speed not file transfer speed.
Oh come off it, that's about as accurate as Telstra spruiking 40+ Mbit NextG with the carefully inserted words "up to"... Something they have been ripped about before even though technically it's accurate.
The problem is most people assume that sync speed = download speed where this is not actually the case.
And the people selling the scheme are doing nothing to disabuse this notion. It's not just the whole 'honesty in advertising' thing (and the innate hypocrisy of holding one organisation to a higher standard than another) that get's in my craw, it's the mismanagement of perception. You create an environment where people believe something, even when those of us in the know understand that it isn't going to turn out that way, you'll end up p#ssing off a lot of people.
Oh come off it, that's about as accurate as Telstra spruiking 40+ Mbit NextG with the carefully inserted words "up to"... Something they have been ripped about before even though technically it's accurate.
There is a huge difference between speed claims of wireless and wired last mile access, and that is that the last mile link is shared in the case of wireless, but dedicated in wired.
Wireless is like everyone in your neighbourhood sharing your driveway – the slightest load would decrease your speed and congestion is pretty much inevitable at moderate loads. Wired last mile is having your own driveway. Congestions on the main roads will still affect your travel time, but you are guaranteed no congestion in your driveway.
But what has this got to do with this thread? No one knows what is causing the fault. They are still investigating. Do you not allow your mechanic to try to diagnose a problem with your car before deciding that it needs a new gasket?
Maybe its been a slow news week. Or maybe there is an agenda for some in the media to try to frame the NBN itself as somehow fatally flawed in technical terms (hint: it isn't).
Simon is being too kind here... it's well known that NewsCorp has a problem with the NBN and with Conroy, piped through the Australian and through other NewsCorp rags, which will stop at nothing to rubbish the NBN with FUD.
Internode is just stuck unfortunately in the middle of all the media muckraking.
There is a huge difference between speed claims of wireless and wired last mile access
And there's a huge difference in actual performance of links regardless of the advertised speed you get.
Telling people to expect lightning fast 100 Mbit connections when, on average, most uploads won't ever run that fast (unless your torrenting or have access to some major upload b'width) is deceptive.
But what has this got to do with this thread?
Because even when this fault is resolved, other people are going to have high expectations which will not be met all the time. Unless someone bothers to educate them of course.
I believe the thread's title starts with "PR Woes?". Seems right on topic to me. = P
Perhaps Simon or another Internode Rep can fill us in on the details then...
We have a staff member on the way to the site today – to work with the school IT staff to go over every bit of their configuration and check it all, and to run our own on-site testing.
Its great that they're happy in the meantime but we're not done yet – we are going on site to fully understand the intersection between performance and expectations at this site.
One of the things this is showing us – and its a hugely valuable learning experience – is that expectation management is a challenge at 100M data rates.
Your network interface might be running at 100M but (as I've said earlier in this thread) there are a host of things that can impact single threaded transfer performance in a 100M environment, many of which are actually instances of all the moving parts working completely correctly (including, for instance, TCP slow-start meaning that a short transfer at 100M speeds doesn't normally hit full line speed before its... done).
Anyway – the point is that we appreciate the school confirming that they're happy now, but we aren't happy yet – and we want to be there to make 100.0% sure there isn't something latent left on the site that might return later on.
Regards,
Simon
Way above ADSL2+ rates, but not magically '100M on demand' right now.
Some of the claims in the media seem to be a bit over-simplified, in looking to declare some specific 'smoking gun' before we've actually had a chance to locate where that smoking gun might actually be.
There's nothing intrinsically wrong with the technology and there's no underlying resource starvation in our backhaul links that we can see... thats the point.
It will be a long time before I can get NBN, but it looks like an Internode customer needs an education – he supposedly works in IT....Heaven help us if that is true...
http://www.themercury.com.au
You go Rodney!