Broadband Report - What do you think?

[Gatecrasher]

A draft report of "Broadband Ratings - The MyADSL guide to South African Broadband Services - March 2005" can be viewed on the following link:

http://home.telkomsa.net/rudolph/march_br.pdf

Basically, broadband services are rated as follows:

Service___________ price - speed - reliability - Support = Total%

ADSL HOME 384_______7 - 8 - 10 - 4 = 73%
ADSL HOME 512_______6 - 9 - 10 - 4 = 72%
ADSL HOME 512 U/Sh__3 - 10 - 10 - 4 = 62%
MY Wireless 256______7 - 4 - 7 - 6 = 62%
My Wireless 128______9 - 2 - 7 - 6 = 54%
Vodacom 3G_________8 - 3 - 9 - 8 = 52%
Iburst______________7 - 3 - 3 - 6 = 41%

The Iburst verdict: Stay away until the service improves.
Pros: None.
Cons: Poor reliability, Speed and Cap.

Is it fair comment? Worse price than Vodacom 3G? :D Slower than MW256? :D Totally unreliable, less than half the score of mywireless, despite virtually no downtime in over 3 months? :confused:

On the whole, my experience has been exceptionally good, and I cannot think of any other SA broadband service I'd rather be using, both on cost and performance grounds.

Personally, I think the writers of the report are living in some strange parallel universe and need to do a lot more research. What do other Ibursters think?

[slimothy]

wtf is that? you can't compare a service that hasn't launched with ones that have been launched for ages, those guys are idiots.

[Gatecrasher]

those guys are idiots.

I wouldn't say that, but I do think they need to select a larger and more representative sample of users/testers. And perform some tests more appropriate for this platform and which suit local conditions. Yes, latency on Iburst is not its most stable feature, but to use a test that treats high latency as an error is inappropriate.

No way will you get 1mbit/s international on a single thread - on any platform. On multithreads, however, Iburst will beat the pants off every other local service. Most people use multithreaded download managers, so why shouldn't multithreaded downloads be a factor in the testing. And why no local bandwidth tests?

Sorry, but I think they may be doing themselves, WBS and potential broadband users a gross disservice if this the final form of their report.

They need to re-examine their results and find out why it differs so remarkably from the actual experiences of many Iburst users on this forum.

[slimothy]

No way will you get 1mbit/s international on a single thread - on any platform. On multithreads,Thats crap, when I'm in SuSe linux I can get 1mbit off a single thread local and international, windows is a diff story though

[Gatecrasher]

windows is a diff story though
so true... so true... :D

[Gatecrasher]

According to the report Iburst download speed is +-120kbps or 15KB/s. So how does one explain the following quotes from this forum:

04-03-2005: Raithlin: downloading a 690MB ISO at an average of 125kB/sec.
05-03-2005: ghostim: i'm getting speeds up to 120 kbytes/sec. Downloaded a knoppix iso in 3 hours
05-03-2005: LandyMan: downloading a 6Mb file from Microsoft at 46KB/sec
06-03-2005: FISHborn: i've got... dl speeds of 120kb. (10mb Telkom test.file in 1:25. And a total dl and browsing of 210mb in 17min
06-03-2005: UnUnOctium: I get full signal when idling and my local and international DL speeds average at +-120 KB/s
08-03-2005: Daveogg: (To the Rodent) I get 116 KB/s so next time you are in a rush for that latest ISO give me a call and i will download it for you on my Iburst.
09-03-2005: crash: I'm downloading CentOS iso files from a UK server and it going at about 60 Kbytes / sec.
10-03-2005: Pan: I get... Your line speed is approximately 943.1 Kbps or 115.6 kBytes/sec
10-03-2005: Slimothy: its not all fun and games downloading at breakneck speeds, I fill up harddrives, run out of things to download etc etc.
12-03-2005: Arthur: I'm getting 800k regularly on the Telkom http speed test
14-03-2005 hentie: Flying at 919kbps as per telkom speed test.
14-03-2005: cAde: it realyu flies!!! , like 120KB/s , signal does do everything
15-03-2005: aborg: 90.4 KB/s local speedtest and 35 KB/s international single thread
15-03-2005: BeLgaRiOn: Been getting fairly constant downloading speeds of 120KB/s. Pretty good
15-03-2005: Swoosh: Happy chappy me!! My connection Just Do it!! like Nike!!
15-03-2005: DaveBuchanan1337: I regularly achieve 0.5mbit on a single threaded international download
17-03-2005 rudids: all my tests are now consistently above 1000 Kbps (international is slightly slower but not significantly), with no latency issues whatsoever:
17-03-2005: BillyBoy: Just downloaded at 87 KBytes/sec.
17-03-2005: Swoosh: doing 1028kb/s at the moment.
17-03-2005: limnos: Connected to Indaba Hotel (5.16km) and i get 900-1000KBp/s
17-03-2005: scavern: Plumstead tower is pretty good as well. I get speeds of up to 120KB/s average is about 90KB/s
17-03-2005: jpd: I have been using iBurst in Cape Town for 3 weeks now - connected to the Belville tower - speeds are averaging between 80 - 100 KB/s
17-03-2005: Vowthorn: getting speeds of about 100KB/s which um... makes sentech look stupid...

The fact of the matter is that Iburst is way faster than any other SA broadband service. Come on, rpm, you really need to find out what went wrong with your research, and fix it.

I saw a thread in the Sentech forum today. Those guys are telling each other not to complain because their horrible 4-6KB/s service is better than Iburst .... and wait for it.... its "official", because it's in your report !?! Yeah, right!

[jpd]

I agree 100% with you - I have used both ADSL and am also a 3G user and iBurst beats the pants off both of these. The authors of the report are misinformed and have clearly not actually done any substantial testing over a period of time. They also fail to mention how easy and quick it is to get iBurst up and running. No red tape, you can walk into the reseller's shop, pick up the equipment right there, plug it in and it works.

ADSL - if they service your area and assuming you don't have fibre running to your premises you are still looking at weeks before you get the line installed etc.

3G - sure you can walk into Vodacom SP store and sign your life away but then u have to wait hours before your service is activated.

[JayT]

stop complaining and next time make sure you and all your friends take part in the tests. Also, make sure you do a few tests a day as once a day when you and everyone else isnt using the connection doesnt give accurate results.

[stepper]

Gatecrasher, wassup?
I understand what u r saying but then its our fault we iBursters we didn't heed (sp?) the call to testings that were announced by RPM. Maybe next time all of us (iBursters) who have the problem with the report we will do our best to 'show' how good is our iBurst.
Ok?

[TheRoDent]

You do realize that the MyADSL report is based on am empirical electronic test, done across all the providers right?

[bb_matt]

I don't think they do Rodent.

In fact, I think they were totally unaware that rpm asked for people to run the bandwidth tests on which part of the report is based.

Yep, we were all asked to participate - did you slimothy or gatecrasher ?

[rpm]

Hi Gatecrasher and others

I can appreciate that somebody with an iBurst connection that is working well does not find the broadband report information to be accurate. Unfortunately we do not have any other information apart from the testing results, and it will indeed be completely unjustified if we ever use information other than that. I am confident that the testing data is quite accurate (exactly the same methods were used for all services), and it is indeed the same methods etc. that placed iBurst on top of the wireless offerings the previous time. The previous time there were no complaints, but obviously the service has deteriorated since then. Many of the same testers, far worse results.

We will adapt the report to indicate that this is a soft-launch service etc., but the results will stay the same. If you have additional comments please let us know. Your input is of great value as usual.

Regards,

RPM

[Gimli]

I would like to see the demographics of the test sample, how many people participated and to which towers they connected to. People have a tendancy to participate if they are unhappy. So if all the unhappy ones on the effected tower participated.... you get a bad result. Next time as you all say we need to get a more representitive sample.

Once this report is released officially it will be damaging to iBurst (something they have to take blame for because there has been some issues). I have seen i.e. Telkom quoting the results of this report in the past in advertisements.

[martin]

I personally know two people who's experiences with iBurst very closely match the findings in the broadband report. I'm not at all surprised with the rating it received.

[Vio]

People have a tendancy to participate if they are unhappy.

Then there must just be many more unhappy ADSL/MyWireless users right? :p

[slimothy]

the problem is that the broadband report was done with software available for windows, so its not suprising that off single thread tests on windows it came up short, but I do think if software was provided for linux as well next time or there was a web test that stored the result data in a database (not hard at all to do) that iburst would come up significantly faster but definatly lose in the ping time arena.

[TheRoDent]

Ahem. What OS do you think 80% of .za Internet users use? Besides, multithreaded downloads isn't something you should have to resort to in order to get decent performance.

[slimothy]

yes but linux performance is better, i mean theres no doubt, so the people that do use linux that would have evened the score out were not included, what about mac users too. You can't go ahead and say you're doing a broadband test but only on the largest demographic. All that coupled with the fact that the service hasn't launched and during the first week of march when the tests were going on we had basestation upgrades and stuff. You need a neutral test (web based), it needs to include only launched services. I hope theres another one in april.

[bb_matt]

It's a guide slimothy, "The MyADSL guide to South African broadband services"

rpm could leave iBurst out, but that wouldn't help people trying to decide what service is best for them.

Linux is not an issue here, as Rodent indicates.

We may as well ask people to indicate what firewall software they are using, or what browser they used for the web based tests in case they "effect the results" ;)

[hArTh]

Slimothy and gatecrasher

Did you take part in the tests? If not, quit your whining.

Your non-constructive whining and calling the people who run MyBroadband "idiots" makes YOU look like an idiot.

[slimothy]

well they were upgrading the tower so I got to participate in 1 day of testing, well when I say one day I mean 2 tests basically. and linux is an issue, unless someone calls it 'the myadsl guide to broadband for windows xp for the first week of march only'

[hArTh]

well they were upgrading the tower so I got to participate in 1 day of testing, well when I say one day I mean 2 tests basically. and linux is an issue, unless someone calls it 'the myadsl guide to broadband for windows xp for the first week of march only'

Wow ... you really are an idiot. You seem unable to grasp the purpose of this ongoing testing at all.

[slimothy]

we all know the score right, my iburst is way faster than your adsl, and your adsl has better latency (woopdeedoo), so all i am saying (in a constructive spirit) is the test needs to be platform independant and if you include services not launch and therefor not as they will be, then at least do it when there are not heavy upgrades going on. I mean it is common knowledge that at the beginning of march every base station here in GP was upgraded then everyone complained cuz the upgrades hampered performance, as soon as the upgrades were done, everyone was saying how good it was, well where were the tests then? so maybe that needs to be reflected in the test results, hah homeadsl384

[Gimli]

Another point regarding testing between ADSL and Wireless devices:

Wireless products are highly reliant on the quality of your connection, or put in another way, your signal strength. You can not take into account people with low connection signal. It is not the fault of the service provider, it is just a function of where the tester is relative to the tower.

Fixed line services does not have this problem.

I would venture to say that the only way to accurately do these tests is for the compilers of this report to properly setup their own IBurst modems using a random sample of towers. Signal strength should be at least between 70 - 100%. And then carry out the tests. Or if you want volunteers, the signal has to be taken into account during the statistical analysis i.e. a test done with a lower signal carries a lower weight.

[ic]

Ag shame, GC & Slimbo - you didn't bother participating, so now you're complaining that the test results don't represent your own personal experience, the point is that the test results do respresent the experiences of iBurst customers that have been stuck with the Bryanston & Northcliff base-stations, which WBS have told us are used by loads of iBursters - I won't venture so say a majority but it could well be the case.

So, how much more representative are results from Bryanston-BS & Northcliff-BS than the undersubscribed (according to Slimbo) Kempton-BS, and GC you're on Midrand-BS aren't you? Would you prefer that tests only be conducted on Kempton-BS & Midrand-BS, whilst totally ignoring the existence of Bryanston-BS & Northcliff-BS?

My advice to you next time is: make an effort to participate in the testing so that we don't have to put up with misplaced loyalty to a service provider.

[slimothy]

ah man dont get it twisted, i'm not saying 'boohoo it never matched up to my tests, therfor its crap' I am sayign I see fundemental flaws in the test and as such I dont think the results are accurate. and for the last time I DID PARTICIPATE, but it doesn't matter if i did or didn't, it has nothing to do with that, I am saying that if one is going to publish these test scores either tell people that they should note a few things, liek the fact that it was windows only, like the fact that some of the services listed are not final. I think its less of a "test" as i think of one in scientific terms and more of a 'you know sort of kinda general test thingy mabob'. I'm not sayign all this to be an ass, I do also own adsl, so I am not invested in any broadband service to the point where I would 'fight to the death for it'. I just doubt the accuracy, and so does my 112KB/s single thread download

[slimothy]

and IC, i would go as far as to say the bryanston/northcliff crap speed users are now in the minority and you should also note that kempton isn't that empty anymore.

[Gatecrasher]

I don't think they do Rodent.

In fact, I think they were totally unaware that rpm asked for people to run the bandwidth tests on which part of the report is based.

Yep, we were all asked to participate - did you slimothy or gatecrasher ?

Well, I was not personally asked to participate. I knew nothing about the testing until it was already under way. Obviously, the request was not posted on the Iburst forum, but in the general section, which I rarely used to visit, and perhaps many new people to the Iburst forum don't visit either.

And although I tried to participate, I didn't get a questionaiore, and the BBST test did not work at all in Iburst, even when the connection was flying on other tests. And the other point is, Iburst really shines on multithreaded downloads, but all the testing is single thread. So I am questioning whether the tests are appropriate. There is also no local testing, which is again an area in which Iburst wipes the floor with the competition.

The fact remains, that on speed, the report is just plain wrong in its conclusions. Now you may argue, that those were the results of those who participated, BUT then it must be stated that only 8 users participated, and that there are many users who are experiencing full 1024/384 speeds.

rpm, regardless of "following procedures" you have to put substance over form, and paint an honest picture. The reality is NOT what you have written. And that should concern you far more than it appears to. Right now your report is an exercise in misinformation. The quotes I plucked above are from 21 different users, not including myself. Are they all bull****ting?

Seriously, you are opening yourself up to ridicule if you proceed with the report in its current form. Your sample size is too small, and is not representative of Iburst users. Any pretence at science in your report is purely dillusional.

[slimothy]

the adsl users are going to kick you in the balls about that multithreading comment, i forsee soemthing like "nooo nooo multithread doesn't count nya nye", and even when in windows my single thread will beat the pants off 512 adsl, and i dont even think i'm in the minority. You know what, I got some time free today, i'ma make a web test

[hArTh]

blah blah blah

Kindly post a scan of your degree in mathematics ... pronto.

I love watching Ph.D's in internet pron lecturing physicists and mathematicians about "science" and "scientific method".

[AcidRaZor]

Die Hippies!!!

[slimothy]

yeah, this is why I hate when ADSL users come to the iburst forum and start getting all 'flamey', they're a sad little bunch

[ic]

...I mean it is common knowledge that at the beginning of march every base station here in GP was upgraded then everyone complained cuz the upgrades hampered performance, as soon as the upgrades were done, everyone was saying how good it was, well where were the tests then? so maybe that needs to be reflected in the test results, hah homeadsl384The reason why iBurst is so unreliable & unstable is bcos this & that & other stuff you're never told about is constantly being messed with by WBS - how do you think Bryanston-BS & Northcliff-BS got stuffed in the 1st place - lack of change control & proper management of changes being made. Accept what RPM has told you, the report will reflect that iBurst was not yet officially launched.

Where do you get "everyone was saying how good it was" from? I don't see 2500..3000 iBursters having posted any such statements...

[Gimli]

BUT then it must be stated that only 8 users participated,

I have asked for the demographics of this test in this forum and we have not had any. Gatecrasher, if what you say is true (and I would like to know where you got the info), I will not loose a single breath over this result. Eight users on a pre-launch service, where probably the most of the participants logged onto the Bryanstone tower (that had problems at the time) is LAUGHIBLE.

Please RPM if this is the truth don’t even include IBurst in the report if you want it to be in any way credible.

I have read somewhere else that MyWireless 512 have not been included as there were not enough participants? Please tell me what constitutes a ‘credible sample’.

[Vio]

Instead of whining about how your iBurst is so superior and it should get more score, rather reflect on the fact that the exact same test was performed on all other services. This is a broadband comparison test. Not a Windows/Linux/Fixed line/Wireless comparison, a broadband comparison.

If the people that choose to participate in testing the system give poor results then the report WILL show the results regardless of whether or not YOUR connection performs fine.

Saying things like "oooo noooo iBurst is in pre-release it cant be compared!" or "Wireless technolagies always have high latensies!" is pointless, you had your opportunity to participate but it slipped by… I’m sure you will not make the mistake twice. If this is damaging to iBrust then so be it since the majority of users for their service that DID participate had a crap experience. :mad:

[Vio]

I have asked for the demographics of this test in this forum and we have not had any. Gatecrasher, if what you say is true (and I would like to know where you got the info), I will not loose a single breath over this result. Eight users on a pre-launch service, where probably the most of the participants logged onto the Bryanstone tower (that had problems at the time) is LAUGHIBLE.

Please RPM if this is the truth don’t even include IBurst in the report if you want it to be in any way credible.

I have read somewhere else that MyWireless 512 have not been included as there were not enough participants? Please tell me what constitutes a ‘credible sample’.

You are looking for demographics to discredit the result i'm sure you'll find none. :mad:

[Vio]

You will find that UUNET uncapped ADSL is not present in the report as there weren’t enough testers...

Makes you think no? :rolleyes:

[AcidRaZor]

makes me wonder about boobies...

[slimothy]

what abotu 3g, maybe i can get better speeds with 3g direct to phone than I can through windows, this test is far from perfect. To be honest I dont care who comes out tops in the test, you think I want more users on my network?? no I do not. If they did the same test again under more scientific principles and it yielded similar results then I would accept it, but right now its just not there. Anyways scientests do something called levergaing in tests, if i race a 3 legged dog and a 4 legged dog, in a scientific test you make calculations to estimate the speed of the four legged dog has he has 3 legs, then compare the results.

main points:
single thread only
windows only
3 mature services against one not mature service
testing from uncontrolled servers

I'm not saying iburst is superior, i am saying that the test doesn't hold any weight, anyways we all know the difference between the broadband services wether theres a test or not, adsl better ping iburst better speed,

[Gatecrasher]

Maybe rpm can post all the Iburst testing results on this thread, and we can draw our own conclusions. And maybe as many Iburst users as possible can post their results on this thread.

This is nothing to do with loyalty, I get miffed whenever I read anything where falshoods are paraded as facts. It is the plain simple fact that what is written in the report is FALSE. WRONG. NOT RIGHT. INCORRECT. Any writer of a report must be concerned when their conclusions are challenged, and when the evidence of how wrong they are is so clearly visable and apparent. Surely, they must question and re-evaluate their methods and not continue blindly onwards.

At least the report should say that there are many happy users getting speeds up to 1 MB/s, or that good results depend on signal/location. Also that there is trial period to test out your signal/connection.

The section in the report on IBurst has absolutley no balance in it, whatsoever. It is so bad and shoddy that it reflects far worse on the writers of the report than it does on the product they are describing.

You can dismiss what I am saying as blah, blah, blah, but at the end of the day, the report achieves none of its objectives if it misinforms.

[rpm]

rpm, regardless of "following procedures" you have to put substance over form, and paint an honest picture. The reality is NOT what you have written. And that should concern you far more than it appears to. Right now your report is an exercise in misinformation. The quotes I plucked above are from 21 different users, not including myself. Are they all bull****ting?

Hi Gatecrasher

I have mentioned in the announcement forum that it will be great to have massive samples where there are users from every tower on every service (including all MW and V3G users), but this is not possible. I think our procedures and testing methods (BBST, TCPIQ and Web Based) are adequate and that a sample of eight steady testers is not bad.

Gimli: I did not gather demographic information (as stated previously), but will consider it for the next testing phase.


Now a few questions to all the skeptics:

1. Why on earth do you believe that your subjective view on the service is more accurate that a week long testing phase that involved numerous testers from all services, conducting various tests and completing a customer satisfaction questionnaire?

2. The report is a comparative report. What methods did you use to ensure that your comparison of iBurst with the rest is accurate, comprehensive and trustworthy? What do you base your comments that it is better on?

3. Why did you not complain about the previous report’s results (when iBurst received 70% and were the best wireless service)? We have used improved methods, and many of the same testers. Their results were worse now than then (especially with TCPIQ and web based: two of the previous testing methods).

4. And: If you want to compare broadband services, please explain how you will do it (practical guide which we can follow). How will you ensure that all towers are covered etc.

Regards,

RPM

[Vio]

Internet providers should know better than limiting speeds on single threads. Or only giving 500kB of broad band. If the speed of a single thread is crap then there is a problem with the service i.m.o., you shouldn't have to have 10 threads open for a single transfer to achieve acceptable levels of speed with any service.

[AcidRaZor]

and what would you do with TWO lesbians, nevermind one....

:eek:

how would you compare them?! boob size? IQ? how good they are in bed? how fast they can make you orgasm?

The point I'm trying to make is this :

Comparing broadband services with each other needs to have seperate variables (like boob size/iq/good in bed etc) to which you can compare the GLOBAL vars to.

for instance speed attained, that is a very common factor.

Now comparing common factors with each other WITHOUT comparing the uncommon factors (like the fact it is wireless etc) would give a wrong indication at the end of the day as to what is the best and what is the worst.

So, having a global comparison between services, then having a seperate comparison between land lines and wireless services and THEN factoring and overall calculation based on those facts, will you know.

For instance, I will include the average distance from tower and average amount of unique towers tested in the wireless section.

rpm? am I wrong? do you already do this?

[slimothy]

first off, gatecrasher, amen brother

then to rpm, it seems simple to me, you set up a simple webtest where users can answer a few questions, so the first one being which service are you running, then i click 'iburst' and it takes me to a page where there are iburst specific questions, signal strenth, location, then finally the web test is run, all results are stored in a database and results can be browsed in real time with all the variables displayed, like number of users that have run the test, avergages, best station, worst station etc. The same could be done for adsl seeing as how adsl degrades over distance, also sentech because peopel in durban seem to have a much better experience with sentech than say JHB, then have the test avaiable all the time to gather information over and extended period of time so we get a better overall picture rather than just one week.

[slimothy]

Internet providers should know better than limiting speeds on single threads. Or only giving 500kB of broad band. If the speed of a single thread is crap then there is a problem with the service i.m.o., you shouldn't have to have 10 threads open for a single transfer to achieve acceptable levels of speed with any service.they dont limit single threads, you dont have to open 10 threads to get 120, and unless you have a digninet no one offers 500KB of brodband in SA

[Gatecrasher]

Internet providers should know better than limiting speeds on single threads. Or only giving 500kB of broad band. If the speed of a single thread is crap then there is a problem with the service i.m.o., you shouldn't have to have 10 threads open for a single transfer to achieve acceptable levels of speed with any service.

Except that multithreading is how many users experience the internet, especially when they intend getting the most out of their service.

If you are going to paint a picture, paint the whole picture.

[rpm]

first off, gatecrasher, amen brother

then to rpm, it seems simple to me, you set up a simple webtest where users can answer a few questions, so the first one being which service are you running, then i click 'iburst' and it takes me to a page where there are iburst specific questions, signal strenth, location, then finally the web test is run, all results are stored in a database and results can be browsed in real time with all the variables displayed, like number of users that have run the test, avergages, best station, worst station etc. The same could be done for adsl seeing as how adsl degrades over distance, also sentech because peopel in durban seem to have a much better experience with sentech than say JHB, then have the test avaiable all the time to gather information over and extended period of time so we get a better overall picture rather than just one week.
Hi Slimothy

How will you ensure one company does not use their 33.6 connections to produce inaccurate results? It is also still only one test…there should be more than one to ensure accurate results.

How about the other three question?

Regards,

RPM

[Gatecrasher]

rpm, I know my test results, although patchy were 2-3 times better than your "average" Iburst speed. TCPIQ was 250-350kbps. And in the BBST I was getting over 500kbps.

And I'm sure Slimothy's were greater too.

Your average was 120Kbps

My question is: Were individual testers given equal weight, or was it weighted by the number of submitted tests? If someone with a poor connection was getting bad results and submitted results every 10 minutes or so, could that have overwhelmed the results of others?

How was it weighted?

[rpm]

Hi Gatecrasher

Please answer the 4 questions. I think that will clear up some issues.

Regards,

RPM

[slimothy]

well you match the users current IP against valid Ips for that service, so iburst... will be 196.46.*.*, you can have a million people on GPRS or a 14.4 modem even saying they are testing with iburst but unless they can match the IP rules thier results are disgarded.

And RPM how did you know some company didn't get peopel on 28.8 modems to run the current test and influence results...

as for the other questions, I was only ever aware of this test, if you say iburst got 70% on the last test I would say those results are probably flawed too. The other questions look rhetoric so I won't go answer those.

Also I just made a thread to do a head count of peopel getting 1mbit or close, its only been open for a little bit and there are already 4 names in it. wait till the end of the day and I bet theres more than 8 for sure, yet only 8 people participated in this test.

[Vio]

they dont limit single threads, you dont have to open 10 threads to get 120, and unless you have a digninet no one offers 500KB of brodband in SA

I was refering to MyWireless's 500 KB's of high speed before slowing down to a crawl. They do this to fool speed tests which typicaly would download a 500KB file to calculate your speed.

[ic]

...
3 mature services against one not mature service
...So, what should potential readers of the report conclude? - simply that iBurst IS an immature service, and that is fair & perhaps after WBS have launched iBurst it will be more stable & reliable, which up to now it has not been for quite a lot of iBurst customers.

I'm tired of all this misplaced loyalty towards WBS & iBurst, my last post on this thread.

[rpm]

well you match the users current IP against valid Ips for that service, so iburst... will be 196.46.*.*, you can have a million people on GPRS or a 14.4 modem even saying they are testing with iburst but unless they can match the IP rules thier results are disgarded.

And RPM how did you know some company didn't get peopel on 28.8 modems to run the current test and influence results...

as for the other questions, I was only ever aware of this test, if you say iburst got 70% on the last test I would say those results are probably flawed too. The other questions look rhetoric so I won't go answer those.

Also I just made a thread to do a head count of peopel getting 1mbit or close, its only been open for a little bit and there are already 4 names in it. wait till the end of the day and I bet theres more than 8 for sure, yet only 8 people participated in this test.
Hi Slimothy

This is a comparative test, so the speeds/happiness of iBurst users is only of value if it is accurately compared to that of the other services. Therefore my question: Why are you so certain that iBurst is better than the rest? Why do you so vehemently defend iBurst unless you have done a comparative test yourself? Why do you feel that your subjective view is more accurate than our full testing procedure?

Regards,

RPM

[hArTh]

There is a simple way to solve this problem. Until slimothy, GC et al produce a *working* alternative testing system which they will then hand over, turn-key, to RPM I will be perfectly happy with the existing testing methodology.

Armchair pundits are two a penny.

PS: I'm not holding my breath ...

[slimothy]

I'm not defending iBurst, twice in this thread I have stated I DO NOT CARE WHERE IBURST PLACE IN THIS OR ANY TEST.

What I do care about is the quality of information I recieve to keep me informed, and so should you, you may as well not even have a test if you just want sort of iffy results.

So basically... you're saying rather than trying to make your test better... I should shut up unless I make my own test.... well Ok then

[slimothy]

There is a simple way to solve this problem. Until slimothy, GC et al produce a *working* alternative testing system which they will then hand over, turn-key, to RPM I will be perfectly happy with the existing testing methodology.

Armchair pundits are two a penny.

PS: I'm not holding my breath ...Actually I PMed RPM stating that I do have code that with a few adjustments could be used, if he wants it

[AcidRaZor]

what has your breathing to do with the comparitive testing?

boobies

[rpm]

Hi Harth

Agreed! That is why I asked the fourth question. In order for us to change our system there must be a very good reason (much better that ‘I feel iBurst is better than the rest because I am happy with it’) and an improved alternative system should be in place. I must say that we have done our best to ensure accurate results, and truly felt completely indifferent about who won and who lost. The results were as much a surprise to us that to everybody else. I have great confidence in the current system, specifically because of its independent, objective nature.

Noone: The wireless services do get bonus points for being mobile, but only in accordance to its respective weight.

Regards,

RPM

[slimothy]

so linux and mac users will not be able to participate in future tests? and you will only have closed tests available to a few people who apply?

[Bishop]

Wow, I am late in this thread.

RPM, the test is run on a regular basis right? I know this round started just before I had to go overseas, so I could not take part, but I know at that stage iBurst was at it's worst for me. My results would probably reflect the report.

If I had to run it right now, it would be a lot better. I am quite happy to accept the results, with the knowledge that the next round will probably see iBurst rising through the ranks again. And hopefully stay up from there on...

[rpm]

I'm not defending iBurst, twice in this thread I have stated I DO NOT CARE WHERE IBURST PLACE IN THIS OR ANY TEST.

What I do care about is the quality of information I recieve to keep me informed, and so should you, you may as well not even have a test if you just want sort of iffy results.

So basically... you're saying rather than trying to make your test better... I should shut up unless I make my own test.... well Ok then
Hi Slimothy

My question again: what do you base the statement on that the results are ‘iffy’. How do you know that our results are not 100% accurate unless you have done comparative testing that produced conflicting results? You can’t just ask people (subjective)…let them conduct a week long test, compile the results and see whether it corresponds with ours (obviously you must do it for other users as well…ie mywireless, adsl, 3g). And then we have our testing system again :D

Therefore, lets work together to incorporate your ideas and testing methods to ensure even more accurate results for the next testing phase.

Regards,

RPM

[slimothy]

ok my little thread to find at least 8 iburst users are happy (happy being quantified by 1mbit or close) has paid off, there are now 8 names in that thread, and only 8 people participated in the comparitive test, so had the test been open and platform independant that would have been reflected.

The test could be awesome if it was to provide area specific information,

I just want to state that I am not trying to go against the grain because I am unhappy with myadsl or because I want iburst first, i just want tests run with more users, over longer periods with more specific information gathered, I think that could help someone in say Sandton see that adsl is better for them, and perhaps someone on the east would be better suited to iburst etc etc,

[rpm]

Hi Bishop

Thanks for the feedback.

Two separate tests were conducted at regular intervals over a one week period (and then one web based test to check that the results correspond). We also checked that there were no significant differences between various testing method results…standard research procedure. I hope iBurst will indeed score higher next time…we need true competition to ADSL.

Regards,

RPM

[Bishop]

rpm,

thanks, I hope you send me the invite mail again next time, even though I missed the second test :)

If I can make one suggestion then, maybe announce the next round within each sub-forum as a sticky? I know that I never read the main forums, I have the iBurst forum bookmarked and just come straight in here. If others do the same, they will probably miss the start of the testing.

[rpm]

rpm,

thanks, I hope you send me the invite mail again next time, even though I missed the second test :)

If I can make one suggestion then, maybe announce the next round within each sub-forum as a sticky? I know that I never read the main forums, I have the iBurst forum bookmarked and just come straight in here. If others do the same, they will probably miss the start of the testing.
Will do! After this discussion Gatecrasher and Slimothy might just make me sign an agreement that I will have to do it :D

[rpm]

Hi Slimothy

I agree that the results can be improved with a longer testing period, more users and a better sample, but since we trust on volunteers it is not always possible. It is like an election poll to predict results: a small group is used to predict ‘census’ results. Our sample does not have an upper limit, and we always invite as many users as possible to join the testing procedure. We even threw in an MP3 player to try to attract additional users! I do however feel that a week long test is adequate. We cannot expect volunteers to test for longer than that!

Regards,

RPM

[Johand]

WTF have Linux or Windows to do with your download speed. Anybody that would see a significant difference between Windows and Linux networking speed should please post conclusive evidence of it. I for the most part think somebody is a Linux fanatic and puts in "Linux is better" into everything they find. I would suggest, that for every "Linux is better" statement, some evidence should be provided.

You should also be more detailed: A Linux machine running only mySQL and Apache with no other services/ daemons except the MUST HAVES, with stable well tested drivers was more stable than an unpatched Windows Server, running every bell and whistle. Or you can make a statement like I would make now: "In my experience, Windows XP, SQL Server and VS.NET was more stable than Fedore 3, GNOME, KDevelop and mySQL." Make certain you seperate opinion from fact.

[Gatecrasher]

1. Why on earth do you believe that your subjective view on the service is more accurate that a week long testing phase that involved numerous testers from all services, conducting various tests and completing a customer satisfaction questionnaire?


I'm not disputing results from other services. My complaint pertains only to Iburst. I can trust my own experience, and I can follow the experience of others by way of this forum. I don't want you to adopt my view, all I want is that you include my view, and that your sample of users/testers is representative. Your results do not tie in with what I experience nor what I read on these forums every day.



2. The report is a comparative report. What methods did you use to ensure that your comparison of iBurst with the rest is accurate, comprehensive and trustworthy? What do you base your comments that it is better on?


I base my comments on my own experience and needs. But I can also read, and I read the experiences (good and bad) of other Iburst users. The problem, rpm, is that you are compiling the report, not me, but if I were compiling the report I would want to make sure that the conclusions reflected the actual experiences of a broad range of Iburst users, and not a tiny sample of unhappy users.



3. Why did you not complain about the previous report’s results (when iBurst received 70% and were the best wireless service)? We have used improved methods, and many of the same testers. Their results were worse now than then (especially with TCPIQ and web based: two of the previous testing methods).


I didn't know there was a previous report until it was mentioned. But why would I argue with 70%? Do you know WHY the results were worse now than then? And are the reasons for the deterioration broadly experienced? Or was it just one or two people?



4. And: If you want to compare broadband services, please explain how you will do it (practical guide which we can follow). How will you ensure that all towers are covered etc.


I think you need a larger sample size, the larger the better. Maybe publicise the testing in the individual forums. It was a long time after arriving here before I ventuired onto other forums. I also think your reports should cover the variability of results (standard deviation). I also think your tests should cover local access and multithreading, after all that is also what users actually experience. And I think it is very important that each user is weighted equally.

[slimothy]

I am not a linux fanboy, so dont go thinking that I am saying 'linux is better than everything'. BUT linux out of the box with ethernet runs single thread downloads faster than windows out of the box over ethernet. I know this because I develop on linux for work and school but use windows as my main desktop OS, if you have read old topics this is apparent. You can tweak windows to give better performance out of single thread downloads to a degree but it still isn't like linux out of the box ON SINGLE THREADS. with multithread they are the same but the test didn't use multithread, so had linux been included, if only 2 people signed up it would have had a significant effect on the results, although iburst would still not have come first but we would have had a more complete picture.

homeadsl384 will always come first, untill adsl192 is included in the test, this is because they promise the least so its easier to deliver

[AcidRaZor]

i could go for a nice cold beer right about now... with frootloops.... and a nice benoni cherry that can have her legs behind her head and do stuff with ping pong balls...

yes... i couldn't care less, and so should you guys. there's better ways to spend your time than argue about how a little electronic signal travels (or does not travel) in comparison with another similar electronic signal.

*puff*

hippies

[slimothy]

if you dont wanna participate dont post saying you dont want to participate, just dont

[kaspaas]

Me thinks that iBurst should take note of this and next time testing is done, encourage their users to participate - especially those who they know are getting excellent service.

Maybe offer another 1GB free bandwidth for voluntary services offered.

If their service is as great as being claimed, I can't think of better and cheaper marketing as "Rated at 99% by MyBroadband"

[martin]

ok my little thread to find at least 8 iburst users are happy (happy being quantified by 1mbit or close) has paid off, there are now 8 names in that thread, and only 8 people participated in the comparitive test, so had the test been open and platform independant that would have been reflected.

....

Actually your thread only proves that 8 people are happy with iBurst and not much more. The Broadband report is more of a technical survey and less of a satisfaction survey. The problem here is that a person may only use iBurst for e-mail and browsing and in that regard be very happy with the product.

Also, the test was open to every one. In regards to the tests been platform independant, I'm sure that RPM would be willing to incorporate this if you can suggest the appropriate tools.

[rabbiddog]

I agree with noone. Time for beer!!!!!

[bb_matt]

Gatecrasher and Slimothy, I've noticed that both of you defend iBurst to such a high degree it's quite sad. You also seem to accuse others of "grouping together" around a specific product - as in "you ADSL users" when your actually doing that yourself.

Talk is easy - rpm is actually DOING SOMETHING as opposed to your bitching and whining about how you see the whole exercise as flawed.

Your the only two people that are bitching - why ? - because your "beloved iBurst" scored the lowest.

The rest of us can appreciate the work that rpm has put into this and the time the individual testers spent sending results to rpm, you evidentally cannot.

[slimothy]

when i say open test i mean, no signing up and downloading chunky software, just a simple web test where you can click and thats it, as for the appropriate tools..... USE HTTP, seriously a simple survey with a speed test, that way you know what the user is on, where they are located and people don't have to wait to see results, they just browse the results real time so they can see only 3 people have tried the iburst test for example and that maybe 2 are from area A and the other from area B. Like i have said before, i have code and you are welcome to it

[JayT]

RPM, perhaps you should include some more web based tests so that the minority linux/mac users can add their 2cents worth.

You could also add to the report that the stats are for the "average" man in the street. Those who are looking for a reasonable 24hr connection to do the basic browsing, emailing and the odd download.

Obviously p2p, multi threaded downloads, large downloads etc were not taken into account and each service will react differently to them. eg the caps, sentechs throttle, quicker multi threaded downloads depending on if the web server has limited speed according to each connection etc

[Gatecrasher]

rpm, I know my test results, although patchy were 2-3 times better than your "average" Iburst speed. TCPIQ was 250-350kbps. And in the BBST I was getting over 500kbps.

And I'm sure Slimothy's were greater too.

Your average was 120Kbps

My question is: Were individual testers given equal weight, or was it weighted by the number of submitted tests? If someone with a poor connection was getting bad results and submitted results every 10 minutes or so, could that have overwhelmed the results of others?

How was it weighted?

Sorry for quoting myself, but this is a very important question. rpm? I really would like to know how weighting produced the +-120kbps speed. Were users weighted equally?

[rpm]

Hi Gatecrasher

Thanks for answering the questions…it clarify some misunderstandings that might have developed during this discussion.

I agree that a larger sample will be advantageous, but this is obviously only possible during the next testing phase. I will definitely advertise it in the separate forums. You will then also have a chance to include your views. Regarding the sample being representative: how does one ensure a representative sample? And then…if the user experiences are so vastly different, is it not a reason why the service should maybe score lower?

I agree that one can gather information from these forums by reading the comments, but for a report we need objective, empirical evidence. Users might be very happy with their service, but it is still much slower or less reliable than ADSL for example. This is a comparative report, which means that even an excellent iBurst service can still finish stone last if the others are better! I think this is one of my main points. Before you can dispute the results, you should be able to show that iBurst is ‘comparatively’ better than the rest. This is not easy without objective testing.

Regarding the deterioration of speeds, reliability etc: We did not concern ourselves with the reasons behind the speeds etc…we were only interested in rating the services. I think the fact that the same users that were happy with iBurst in the first test were unhappy in the second (and showed much lower speeds, reliability, customer satisfaction) gives an indication that something might be wrong. I appreciate that some users experience great service from iBurst, but is this representative of the full iBurst population, and again: much better than 3G or MW?

Maybe we must move from here to see how we can ensure that the next testing phase will have a much larger sample size. With such samples one has many statistical options to ensure very accurate results. I can assure you that my main aim with this report is to produce the most accurate results possible (we might have to defend our formulas and testing methods as part of an academic paper, so this feedback is great!). Any suggestions will certainly be considered and if it is an improvement it will be implemented.

Regards,

RPM

[rpm]

Hi Gatecrasher

The weighing depended on a few factors, including the frequency of the tests and the number of days on which tests were conducted. A results from a person conducting only one test using one application must obviously carry a lower weight than a person that have done numerous tests using all three methods.

Regards,

RPM

[slimothy]

Talk is easy - rpm is actually DOING SOMETHING as opposed to your bitching and whining about how you see the whole exercise as flawed.
I'm not whining, i'm trying to contribute thoughts as to how the test may be improved in the future, what? am i supposed to go open another website and have another test? whats the point, why not just help make this test better.

also we are not bitching because it scored the lowest, look adsl384 gives out a measly 30k or so, any server in the world at any time can give you 30k on a single thread, so when you look at promised speed versus speed recieved on a single thread they are close to the same, so homeadsl384 scores nice and high. But if i had a 10mbit, i won't get 10mbit from another server off a single thread will i? i can fill my pipe with 10mbit, but not off one single request, same applies to iburst

[rpm]

I must be off now...will answer your next posts later :D

[JayT]

And for the linux guys.... install WINE and see if the tests run. RPM, you might want to get some official tests done so that there arent any complaints in the future

Linux: http://sourceforge.net/projects/wine/
Mac: http://sourceforge.net/projects/darwine/

[Vio]

I think you need a larger sample size, the larger the better. Maybe publicise the testing in the individual forums. It was a long time after arriving here before I ventuired onto other forums. I also think your reports should cover the variability of results (standard deviation). I also think your tests should cover local access and multithreading, after all that is also what users actually experience. And I think it is very important that each user is weighted equally.

Some constructive points there, but I’d like to stress that international through put is what counts as most content is hosted offshore.

[Daveogg]

Sorry jumping in here late.
RPM if you really only had eight testers on iburst representing a user base of over 2500 your results are completely meaningless. Anyone doing stats 101 will confirm that for you. You are now using useless and baseless information to influence potential broadband adopters. To maintain your and this forums credibility you have to can it now.

[Gatecrasher]

The weighing depended on a few factors, including the frequency of the tests and the number of days on which tests were conducted. A results from a person conducting only one test using one application must obviously carry a lower weight than a person that have done numerous tests using all three methods.


This reiterates my point. You already have a small sample size of 8 users. But in arriving at 120kbps what is the largest weighting given to a single user. What if, say, 50% weighting is given to just one user with a bad connection who submitted gazillions of horrible test results. Will that not skew your results such a degree that they are completely worthless?

If I get 1Mbps and 7 others get 0kbps, then, with equal weighting, the average is comes out to the result you are publishing

If I get 0kbps and 7 others get 1Mbps, but because I submitted a gazillion tests, I'm get 88% of the weighting, the average also comes out to the result you are publishing.

But the two results are not equal.

[rpm]

Hi Dave

Thanks for the feedback.

I agree that a sample size of 8 is rather small, but if we consider the population of 50000 of ADSL it means that we need a sample size of 400 to be within academic research limits. This means that we will be unable to produce such a report…ever! I think this report has great value and give accurate results. If we can find conclusive proof that this is not the case we will gladly can it. Here I am not talking about ‘I get great speeds’, but rather empirical results from a group of users over an extended period (including the same results from users from other services).

Regards,

RPM

[JayT]

Gatecrasher, there is nothing stopping the others from doing multiple tests. And dont forget, soon the iBurst cap will be in place so I doubt there are going to be people wanting to just waste all their bandwidth for a test. Because then the results will be accurate as they sit on a 64kbit connection for the rest of the month. Accept the results and make sure everyone you know participates in the next test.

Also note that results such as tcpIQ are not totally accurate. If I download an update or busy downloading mail, the results drop below 100kbps. As my tests are scheduled to run every 30min on a different machine, I get a lot of inaccurate data. But its accurate in the sense that my speed is xyz during average use. Obviously when Im a sleep and not using my connection the results 'should' increase.

[rpm]

Hi folks

This thread is going nowhere. I think the best way forward is if each member gives his/her view on:

1. How you think we can improve the testing procedure
2. How we can ensure a larger sample size.

Regards,

RPM

[bb_matt]

1. More people, allow for drop-down boxes to rate price, speed, reliability, support (see point 2)
2. Easier entirely web based tests (no need to download anything or install a plugin)

Where I'm going with this is an easy to use single web page - I'm no programmer, but I do know that the current testing method will put a lot of people off.

The drop downs allow testers to use thier subjective thoughts on the service as a whole, the speed part will obviously be weighted against the actual speed results.

Name :
Service <select from dropdown>
Price : 1 to 10
Speed : 1 to 10
Reliability : 1 to 10
Support : 1 to 10

[Click to test your connection]

[slimothy]

i think thats what i suggested way back on the first few pages, yes bb_matt, thats what we need exactly

[Gatecrasher]

1. More people, allow for drop-down boxes to rate price, speed, reliability, support (see point 2)
2. Easier entirely web based tests (no need to download anything or install a plugin)


bb_matt, I agree completely. But how do we ensure integrity? That people actually have the service being rated... :)

[slimothy]

as i said before you just make sure they are on the right IP block, if a user says they're on iBurst and they have the IP of 196.46.*.* you can be sure they are on iBurst.

[Gatecrasher]

as i said before you just make sure they are on the right IP block, if a user says they're on iBurst and they have the IP of 196.46.*.* you can be sure they are on iBurst.

Good for Iburst, but maybe not so straightforward for other services. Presumably, the same methodology must apply to all.

[Gimli]

Hi folks

This thread is going nowhere. I think the best way forward is if each member gives his/her view on:

1. How you think we can improve the testing procedure
2. How we can ensure a larger sample size.

Regards,

RPM

Correct, you obviously have something to defend (your paper). There are plenty positive suggestions on this thread that you can work into your methods to ensure a better test if you do not just pass it over as 'not scientific' or 'subjective comment'.

Lets take another tack on this, please publish your complete testing method, formulas and algorithms that are so 'complex' as per your report. Also all testing data used. Then we get the chance to scrutinize it and let us then see if it can stand up on its own two feet. Like proper scientists do.

[qDot]

How about doing a protocol by protocol or application by application test such as HTTP | FTP | P2P | etc.

We measure that against each 'broadband' provider.

That way when asked "which is better?"

I can reply "What do you want to use it for ? "

and be able to say well iBurst is the best for browsing ... etc.

I think the tower info and even tower sector can be useful that way you should be able to say well if you are in Sunninhill then browsing on mywireless is better from Sunninghill than browsing from Paulshof (same tower different sectors correct ?)


So a laborious test excerice such as the one i suggest can make it all the more difficult for a technical person to answer the question from a non technical person who wants to know

"WHICH PRODUCT GIVES ME FAST INTERNET?"

[jmn]

Agreed Gimli

I actually get better results from the TCP/IQ test when my connection is under heavy load than when it sits idle :) Contradictory at least...

[Daveogg]

This whole process is just flawed. If you have an axe to grind with Iburst, just take a drive to an area on the very limit of reception. Do all the speed tests post a few pings and complain about the "horrendous" packet loss and inconsistent pings. To accurately asses Iburst you would have to have some way of verifying signal strenght frame rate etc of the tester, otherwise you may just be assesing geography.

Put it another way. When i am at work, Iburst is amazing, last weekend i was watching live Belgium Classic cycling (OT for any cycling fans try www.cycling.tv) streaming in at 350kbps and NO buffering stops.
At home i am lucky to get 10kbps, but i am on the edge of a very pale green zone. I am a realist and never expected it to work here. Why do others sign up in dodgy areas then moan about the speed and pings??

IC dont take this personally i understand that you were rendered a poor signal due to antannae alterations AFTER you signed up. Thats a balls up by iburst and you should be compensated.

[slimothy]

Good for Iburst, but maybe not so straightforward for other services. Presumably, the same methodology must apply to all.they all have specific IP blocks

[Vio]

I believe bandwidth testing is the way to go. "Asking" someone to fill in a form is not a good idea cause Sammy that just checks his mail thinks the service is astounding.

I do think that there should be categories of tests to perform as business / downloaders / gamers all have different views of how a service should perform.

Regarding bandwidth / latency / stability tests, why re-invent the wheel the existing benchmark applications are good enough. Other services use them why shouldn't we?

The only problem is the amount of people performing the benchmark is not necessarily a fair reflection. To resolve this problem why not let MyADSL members specify the broadband connection(s) they use as part of their registration information and send out request emails to participate in the tests when they happen. This will increase the size of the test sample dramatically.

That’s just my 2 Zim cents :D

[bb_matt]

Look guys, this is really going way too far !

I did as noone suggested and went for a beer (good call mate, altho I usually do on a friday ! :D) - but now I'm back and suitably refreshed, it seems glaringly obvious to me that rpm never meant the broadband guide to be taken as any kind of defacto standard !

It's a guide !

It's as subjective as saying eating at Browns in rivonia is better than eating at McDonalds, Greyston drive !

The ONLY people who have a problem here are the iBurst users that are getting fantastic service.

That's fair enough - I can understand why you'd be peeved (to some degree)

As I said earlier in this thread - IT's A GUIDE.

Are you going to get huffy about "The Ultimate Guide to Paris" when you end up getting mugged at a place they indicate is "a safe haven for tourists" ?

Will you write to your local MP when "The Guide to Good Haircuts" directs you to a place that supposed to make you look like the second coming of haircuts, but results in dogs trying to hump your head ?

No, of course not.

So give it a break already and have another drink/smoke/carrot (or all three combined)

And noone, BOOBIES !

[Daveogg]

bb mat. If you were a broadband newbie looking to take the leap onto broadband and were presented with the Official Myadsl Broadband guide would it influence your decision?

Secondly Telkom has already used it too suit its own purposes previously.

So if you are going to do this kind of research and promote it as the official mouthpiece of this forum you had better get it right.

[bb_matt]

bb mat. If you were a broadband newbie looking to take the leap onto broadband and were presented with the Official Myadsl Broadband guide would it influence your decision?

Secondly Telkom has already used it too suit its own purposes previously.

So if you are going to do this kind of research and promote it as the official mouthpiece of this forum you had better get it right.

Oh for goodness sake Dave, rpm is doing the best he can !

No I wouldn't just take the guide as the only area to look at when looking for broadband services...

BUT

It's a DAMN SIDE BETTER than relying on the advertising of all the companies involved in the test !

I'm with the Yes sayers here, not the Nay sayers, who just have absolutely NO vision.

Have you read none of the postive aspects about this guide ?

The fact that it's still in it's infancy, the fact that rpm KNOWS that participation should be higher - for fsck sake, if you really want the guide to be more accurate, get everyone you know to participate !

There's nothing worse than people whining about what is to all intents and purposes, volunteer based ! - Yes, YOU can get involved ! YOU have the say.

It's a no brainer really.

[Daveogg]

bb mat when reading any research one has to answer two questions what is the quality of the research and what are the results. To asses the quality of the research one needs to know about the methodology and any confounding variables. With this report as presented one cannot asses the quality of the research as the methodology is not available.

What is a no brainer is too accept results without questioning their validity.

Now i am not trying to dish RPM, i understand that he is doing the best with the resources available however the process is flawed and if the vision is too acurately compare the available offerings then these flaws need to be debated.

[Gatecrasher]

Children, behave!

106 posts in 18 hours. It might "only" be a guide, but in the context of why we are all here it IS important.

There is a thread on the announcements page seeking to make the report the very centrepiece of the site.

What was the winning slogan? MyADSL: The broadband authority.

But to be an authority you need to have best quality information you can lay your hands on.

rpm put a lot of effort into his "only a guide". So why not make it as authoratitive as possible? Why not make it information people can rely on.

The report as it stands in relation to Iburst is pathetic and clearly either the quality or the weighting of the data collected was sadly lacking. That has the potential to reflect badly on the credibility of the report and also on this site.

Whether that is because far too much weight was given to a vast number of poor results from one or two disgruntled users, or because the overall pool of users was just too small doesn't really matter.

As rpm himself posted on the 14th March, the highest session speeds acheived in the testing was 777kbps, single threaded, international from... Iburst! Where does that appear in his report? Nowhere. And how did that manage to get whittled down do a 120kbps service, slower than 3G and MW256? Sorry, but that just goes way beyond comprehension! :eek:

For the findings to be credible they have to reflect the experience of the average user, and not necessarily the average item of data volunteered.

As the person compiling the report, rpm, and nobody else has the responsibility to ensure that his data collection methods are fair, unbiased and respresentative.

It is not the responsibility of users to ensure their data is included. Imagine if this was an election poll to find the next US president, voters wouldn't be phoning up Gallop to make sure their opinion is included. No, Gallop would be getting opinions from a representative sample. If gallop fails to ensure a representative sample, then its findings are meaningless.

I have a lot of symphathy for rpm. He is doing a great job, and this is not a personal criticism. But on the basis of the tiny sample and the strange weightings given to the data, and the contradictions within his own findings, he is in no position to make the derisory and derogatory statements that he makes in the report in relation to the Iburst product. He is seriously misinforming his readers.

[bb_matt]

Dave, do your own research and then come back and complain.
rpm has been through this with us already and was open from the start.

I'll defend the guy to that last, because he has that thing called Integrity.
He's making a stand, trying to make a difference.
He has the start of a broadband test, developed in his own and others spare time and all you can do is join the chorus of nay-sayers.

What are YOU doing ?

Your hammering his and others efforts.

I'm really saddened by the attitude of a few people in this thread - they are most certainly not team players and nor are they logical thinkers.

I can only conclude they have alterior motives.

[slimothy]

what am I doing? well lets see I am implementing a speed test of my own for iburst users, it is web based so will work on any platform it attempts to gather specific infomation from the user and it can run forever and the statistical information will be calculated live to hopefully give a complete picture of where/when the service is good or bad. it'll be up tomorrow

[bb_matt]

I am implementing a speed test of my own for iburst users

Only for iBurst users ?

So we would only get results from iBurst users based on a test only for iBurst users - now that's useful - nice one !

[Daveogg]

bb mat I am criticising his research not him. I critically review research everyday believe me its important in fact it may one day save your life.

As i said before i would like to open the debate on how this should be done.

A few points for everyone to think on.

What exactly are we comparing here. WBS publish a coverage map implying certain area should expect full speed and other area will experience degraded performance. Would it be fair for me to subscribe to iburst and then conduct tests from a degraded area without that being stated.

Are we comparing the average ibursts experience with other broadband offerings or are we only going to comparing users who test from areas that iburst claim to have full speed coverage?

[Daveogg]

Slim maybe you have a point here. Is it possible to develop a system where a user can log a geographical position and then get a speed test done. We can then tell others thinking of getting iburst where it works and where it is failing.

[bb_matt]

Daveogg, you'll find that it was an open test if you look through the posts in various sections of this forum prior to our current date.

As far as I can see, the iBurst section of this website is the most complaining and angry.

[slimothy]

Slim maybe you have a point here. Is it possible to develop a system where a user can log a geographical position and then get a speed test done. We can then tell others thinking of getting iburst where it works and where it is failing.Yepp thats exactly what I'm aiming for, I will be using the existing maps at iburst.co.za and using my dynamic image skills with php to do some dynamic reporting.

Matt, its iburst only because this is the iburst forum,maybe in the future if I could get some information from you adsl users I could implement somethign with ADSL and then do comparitive studies, but its starting with iburst and the point would be to build an overall picture of iburst in south africa, which is valuable for new users especially but also for people on iBurst now.

I don't appreciate the "now that's useful - nice one !" comment, basically I was told if I had better methods I should share them, so I offered my code, no one took me up on it, then I was asked what I am doing, so I will show you how I would have done it, while making it usefull to people from the iBurst point of view (seeing as thats what I use) and if someone from myadsl see's it and wants the code, they're free to it. If nothign else I believe my little test will prove that the iburst demographic wasn't represented well and that there are alot more happy iBurst subscribers than the pecimists would have us believe.

[bb_matt]

Matt, its iburst only because this is the iburst forum

There is no middle road for you, is there Tim ?

[Gatecrasher]

I'll defend the guy to that last, because he has that thing called Integrity.
He's making a stand, trying to make a difference.
He has the start of a broadband test, developed in his own and others spare time and all you can do is join the chorus of nay-sayers.

He's a great guy, doing a great job. But he's not the pope. He is not an untouchable. He can make mistakes. We can all make mistakes. Just because he is a great guy, doesn't mean we must all surrender our objectivity.


What are YOU doing ?
Your hammering his and others efforts.

Rubbish. This is a healthy debate in which the main aim is to enhance his efforts.


I'm really saddened by the attitude of a few people in this thread - they are most certainly not team players and nor are they logical thinkers. I can only conclude they have alterior motives.

What motives? Spit it out! That is just paranoia beyond belief. Why do always have to bring debates down to personal insults? It detracts from the very good arguments you often make.

[jmn]

WBS publish a coverage map implying certain area should expect full speed and other area will experience degraded performance.

Actually burst rates up to 1 mbps. I wish I could expect full speed full time.

BTW, just had 640 kbps average over a 7 hour period continuous download (download size divided by time, 3 - 6 threads running) plus TCP/IQ test running every 2 hours and general surfing and mail checking during that time.

[slimothy]

i think thats our point exactly, you can't expect 1mbit, but I can expect 1mbit or over anytime of day, thats why we need to draw a better picture as far as iburst is concerned,

[Daveogg]

http://www.tcpiq.com/tcpIQ/LineSpeed/Results/TestStats/?CountryId=196&ConnectionId=11302&UserId=154311

Somewhere in this thread RPM asked us doubters to post our TCPIQ results, well there it is. If you want to save some time Download 544kbps upload 112kbps.

[jmn]

i think thats our point exactly, you can't expect 1mbit, but I can expect 1mbit or over anytime of day, thats why we need to draw a better picture as far as iburst is concerned,

An 'official statement' from WBS as far as reasonable speed averages with reference to the map codings would have been helpful... Or is this perhaps the purpose of the test phase? Along with capacity handling?

[slimothy]

and thats on a single thread

[Daveogg]

Yup and just for fun set flashget to single thread to d/l the proggie. Timed it at 59 sec and its just on 5MB so if my maths is correct that is an ave of 80KB/s

[Gatecrasher]

Well, even rpm says that Iburst was the fastest during the testing, 777kbps.

But he didn't put that in the report. Clearly this data carries very little weight when someone out of signal range is submitting a gazillion +-0kbps tests continously throughout the 7 day testing period.

What rpm does say in his report is: Stay away until the service improves. Who should get it? Only those who cannot get another service.

Unfortunatley this just does not allign with the vast majority of user comments on this forum. For that reason, rpm should be seriously questioning the validity of his data, and the conclusions he has reached.

[AcidRaZor]

Maybe they should drop the "comments" of each service rather? focus on speed tests or whatever they do alone. Then disclose how many people participated in the speed tests (and in cases of wireless connectivity) which area and the average distance from the towers.

Then make it last 3 weeks instead of 1, and publish the report the last week (or first week) of each month.

Just some ideas off the top of my head

[bb_matt]

Other ideas would be to have extended testing over a 1 month period, using a similar method to the last test - and make sure people get involved.

An idea here would be :-

20 iBurst users in a widespread demographic - iow, not all connected to the same towers
20 512k ADSL users
20 384k ADSL users
20 192k ADSL users *is 192 broadband ?*
20 Sentech Advanced (widespread demographic)
20 Sentech Premium (widespread demographic)
20 Vodacom3g users (widespread demographic)

The lowest Sentech package is definately not broadband, so it should not be included.

IOW, the people volunteer and HAVE TO complete a specified amount of tests for thier data to be accepted.

20 users each doesn't sound like much, but I think rpm would be the first to say getting more people involved is damn difficult - 20 of each service listed above is 140 people - that's probably do-able.

[AcidRaZor]

I doubt you will find 20 Sentech premium users.

The reason why I suggested 3 weeks is for "publishing" time. (As I'm now contracting to publishers and learning about the way they do business)

3 weeks, then 1 week to publish, then do the whole thing over again for the next month.

[slimothy]

just turn it into a web based test and I guarantee you get those numbers

[ScrnScrm]

for what its worth. hehehehehehe. my opinion :
i have 512kb/s business adsl uncapped (thorugh ISDSL)
vodacom 3g card
mtn edge via cellphone
am on the iBurst trial.

i think the report is pretty consistent with what i have experienced using all four of these services, although EDGE hasnt appeared in the report yet...
well done guys!

as a side note - i have recently returned from the USA where i had Optimum cable internet . It cost $80/month with an unlimited TV bundle, but i must say that I found the service rather ****. Download speeds during peak where often worse than using a standard v90 modem. Unfortunately earthlink DSL wasnt avail in the area....

later

[slimothy]

ok, well me and the other 20 people who supported the thread to see how many people were getting 1mbit on iburst would disagree with you

[ScrnScrm]

huh? i just said that my experience is consistent with what the report says about the relevant offerings. i wasnt slamming your methodology or desire to do further investigations on the iBurst service (which would be most welcome), merely stating my (very well educated) opinion... I dont think the report was very far off the mark!
by the way, I live 180 meters from the iburst base station, and still havent seen the promised speeds...

[slimothy]

huh? i just said that my experience is consistent with what the report says about the relevant offerings. Yeah I know i read the post.



i wasnt slamming your methodology or desire to do further investigations on the iBurst service (which would be most welcome), merely stating my (very well educated) opinion...you don't have to be too educated to know you are either getting 1mbit or you're not and that it is either great for you or crappy for you.


by the way, I live 180 meters from the iburst base station, and still havent seen the promised speeds...it has nothing to do with distance, it has everything to do with signal strength and quality of signal on top of the fact that if you connect to a problematic tower you're not gonna get full speeds obviously. can you do me a favour and paste your ut firmware version?

[ScrnScrm]

lol. ok, will post the firmware versions tuesday when i get back home. dont keep that info at the top of my head :-) I suspect it may be old because i got one of the first shipments, and lost track of when it was last updated...
wrt the distance statement - i am well aware that its about strength and quality of signal. Me thinks my problem is with the quality of the signal... To be honest - i havent bothered much with it lately - just havent had the time to fiddle. Damn, i am supposed to be testing iBurst for WBS...

[slimothy]

you haven't bothered with it latley.. well there you go. for all you know they could have upgraded you to 100mbit and stuck a basestation up right behind your house for your own personal use.

[ScrnScrm]

for all you know they could have upgraded you to 100mbit and stuck a basestation up right behind your house for your own personal use.

oh bugger. i hope they havent. that would totally ruin my view :D