Surprising Results from my Broadband Test

Here are my results done from 1230hrs to 0115 hrs SG-time.

Using Mozilla browser, I have visited the following websites like STOMP, CNA and Razor.tv and accessing the same clips within the same time frame. And to make it slightly more scientific, I have accessed speedtest.net to run 3 instances of the same test to make sure I have a good spread of readings.

Yes I did plan to do a peak versus non-peak hour comparison but what stopped me was that of the 2 services, i.e. the premium and the 'New Super Fast' service, I would guess that premium would have more customers than ultimate. So to make things 'fair-er', I decided on the sleeping hours of 12 to 1am where the line is much clearer for both services.

To make it as fair as possible, I reset the CSU/DSU and also disable my internet connection every time I hooked up to the CSU/DSU to ensure connections is 'fresh'. Not that it will matter lots, but at least all started from the same condition.

As expected, the video downloads from the local websites are pretty fast with fast downloading of web pages and most importantly the videos. I did try to see how much quicker but in the end, it is just a matter of seconds. Practically speaking, it is a short wait and the differences between the two is really negligible.

Since I can't differentiate the two in terms of the timing, then speedtest.net would be a good indicator of connection speed.

As mentioned before, there will be 3 instances of the same test but will be done on a local server from New Media and a San Francisco server hosted by Unwired.

Here are the results

The Super Fast Service (1st Test/2nd Test/3rd Test)

Local Server (New Media)
Download - 20.59/21.26/20.87 Mbps Avg: 20.9 Mbps
Ping - 32/29/31 ms Avg: 30.6
Upload - 0.73/0.72/0.70 Mbps Avg: 0.71

Overseas Server (Unwired)
Download - 1.22/1.75/1.03 Mbps Avg: 1.3
Ping - 238/233/237 ms Avg: 236
Upload - 0.52/0.52/0.53 Mbps Avg: 0.52

Ultimate Service (aka 12 Mbps service)

Local Server
Download - 23.35/23.4/20.34 Mbps Avg: 22.3
Ping - 28/25/26 ms Avg: 26.3
Upload - 1.10/1.10/1.10 Mbps Avg: 1.10

Overseas Server
Download - 3.74/2.82/2.79 Avg: 3.11
Ping - 204/206/205 Avg: 205
Upload - 0.61/0.62/0.61 Avg: 0.61

Conclusion:

a) Local content - given that the user experience is pretty fast for both services, there is really nothing to differentiate them. Perhaps the server allowing access to the content is being curtailed to a point that it is quite uniform between the two

b) Overseas content - given that most users would travel overseas for their content such as Facebook, this test is pretty significant. But what is the issue with Internet infrastructure is that all the data packets do travel around the world in different routes,through different routers and connections. So any comparison in timing is 'unfair' in the strictest of senses. However speedtest.net somehow give an acceptable results to make a fairly good decision.

So what happened here is that local content is pretty straightforward as it is fast and usable. What surprises me is that my 12Mbps service with an older CSU/DSU is able to show results of average 22.3 Mbps data throughput rate from a local server. That is nearly twice what the older CSU/DSU is capable of! It is really weird. Has my hardware upgraded itself? Or is the test site gone berserk?

My 12 Mbps service has consistently beaten the 'Faster' service for both local and overseas server and even the download speed from overseas server is 3 times faster.

Now I really have to question what in the world is going on? I may have enjoyed a faster speed than what was advertised (great!) but are the people paying higher speed packages really getting their monies' worth?

Judging by the throughput rate of the local server, I would harbour the guess that in the effort to outdo competitors by increasing the download speed, the connection speed is way way above what the servers can 'serve' out. So any increase in speed is really not necessary unless the server that one is accessing can support the speed nearer to the services one has subscribed to. Otherwise it is really not maximising the online time (sorry for the pun)

Ok, let's just say speedtest.net is 'spoiled' and using my nonscientific method of measuring video downloads, the difference is really minute. So I would take any of these 2 services for local content. What's left is to choose the cheaper of the 2.

Rright now I guess I just stick with my old CSU/DSU.