Log in

View Full Version : Vodafone Mobile Broadband - Service Quality Testing



Hands0n
2nd May 2010, 01:44 AM
Much is made about the speed of both broadband and mobile broadband as if that is the be-all and end-all of it. But that is far from the case. The actual quality of the service you receive is really much more important than just raw speed.

And so it was pleasing to see that the good folks over at speedtest.net have now got a new service on the go called Pingtest.net (http://www.pingtest.net/).

With pingtest.net you are able to take a measure of the actual quality of the broadband service that you are receiving and, like with speedtest.net, whether that service level is consistent or not.

The tests below are taken within minutes of each other at 1:33am when the local transmitter should be little used.

At this time the speedtest result was a respectable one for being out in Kent. The local 3G Vodafone mast is about 200m away from where I am sitting :)
1098


Next I ran two sets of pingtests with the results below. What these reveal is that the test has resulted in a measure of service giving a Class "C" line quality. This is suitable for the majority of activities, but would not be good for gamers (high latency, ping times) and Jitter that would impact VoIP communications. A Class C service may also not be very suited for streaming video from the likes of the BBC iPlayer.

The two tests were very consistent but would need to be taken at several different times of day to produce an overall picture of what the quality of service is like.

Test #1
1096

Test #2
1097

Why is Jitter so important? Basically, Jitter is a measure of the variation in time between data packets arriving at their destination. Often this will be caused by variable congestion along the path that these packets typically take. This may be caused by your local facility being shared with other users at the same time such as the airtime or backhaul circuit associated with the local 3G mast. It may also be caused by conditions deeper within the Internet itself.


... jitter is the variation in the time between packets arriving, caused by network congestion, timing drift, or route changes.

A good read on Jitter as applicable to VoIP is here http://www.voiptroubleshooter.com/indepth/jittersources.html - although the concepts apply equally to any other type of data regardless of whether it is critical or not.

Either way, Jitter is disruptive to time-critical data such as voice and video.

And so the above pingtests reveal that the Vodafone mobile broadband service quality is fine for general purpose Internet use. It may, however, not support very well on-line gaming, VoIP (i.e. Skype) or streaming video. It is unlikely that any other mobile broadband provider will be able to better Vodafone as it is the 3G or UMTS technology that is the limiter here. Our next generation networks (LTE or Long Term Evolution) may be better suited, but these are still a few years off from being generally available.

DBMandrake
2nd May 2010, 09:32 AM
Yes pingtest.net is a very handy site, I've known about it for a few months, and it reveals one of the statistics about internet connections which is not widely reported or measured...(eg jitter)

I don't agree with the suggestion that other networks aren't able to better Vodafone's results however, as there is a lot of tuning that can be done in a 3G network that is optimized for data to reduce jitter - and frame times in 3G are 2ms so in theory the minimum jitter is limited to 2ms in ideal circumstances.

I thought I'd test my iPhone 3GS on 3 - Performed around 9am this morning, (the results pages say 8am but they don't seem to be taking into account DST!) bearing in mind that the iPhone does not support HSUPA, which limits upstream speed to around 300-400kbit, and also increases latency by a few 10's of ms compared to a phone with HSUPA.

I'm about half a mile, eg 800 metres from the nearest 3 mast.

I'll let the results speak for themselves :)

1099

1100

1101

Hands0n
2nd May 2010, 12:25 PM
The low Jitter that you have seen on 3 are creditable indeed, and probably a consequence of 3's network being built out in 3G/UMTS from the off. There has been no need to shoehorn in the new technology onto an incumbent data network. The "big four", however, have had to do just that. Combining 2G, EDGE and 3G into their Intranet and no doubt making a right hash of things with firewalls and whatlike. A traceroute is often very revealing of the internal complexity of their network - and Vodafone does seem to have a lot of internal hopping around before the data leaves their network.

It would be nice to see a set of results from each network posted into this thread for comparison purposes :)

DBMandrake
2nd May 2010, 12:57 PM
Yes it does seem likely that trying to shoehorn data into networks that were originally designed for voice won't give as good results as a network designed with 3G data in mind from the get go...some of the networks probably still have a lot of legacy infrastructure that they're unwilling to retire until it's paid for itself...

I tried the test again on the same server just now at lunch time and I was getting 17-20ms jitter, so I thought "opps", but then I tried Milton Keynes and jitter was 8-9ms so I think that shows that a significant part of the jitter can be the test server (and/or the connection across the internet to get to it) so not all the jitter is the 3G network itself.

I would like to test the other networks too, (and have SIMs for all of them) unfortunately I can't enable tethering on the iPhone on any of the official iPhone networks as they are locked down by the carrier settings bundles :( The only other network I can test is T-Mobile, and because T-Mobile throttle speeds on basic accounts I can only test it in a meaningful way by using a Virgin SIM, (as they no longer throttle speeds) so here are the results:

1104

1105

1106

Jitter and latency also commendably low, slightly lower than 3 at the same time of day, but the speed is pretty poor indeed at only 0.48Mbit and that was the best out of a run of 3 attempts. Just to check it wasn't the different time of day being unfair on T-Mobile I popped my 3 SIM back in and retested it and I got 1.2Mbit.

Although 3 and T-Mobile have merged their 3G access networks in many parts of the country they have not yet to my knowledge merged within my coverage area, and still have separate masts about 500 metres apart...(with the T-Mobile mast being around 300 metres from me, and the 3 mast 800 metres away)

Also interesting that the funky firewalling that T-Mobile and Virgin do is preventing the packet loss test from working...