For a brief period of time, I will be having both Singnet and Starhub broadband services in my home, so I thought it would be fun to run speed tests and see if I’ve made the right decision in switching (I shan’t say from which to which).
Both are 8Mbps plans, with Singnet promising up to 512kbps of upload bandwidth, and Starhub
only half of that, at 256kbps up to 768kbps after a recent free upgrade. Their prices are comparable. I connect to both via Wi-Fi, although my Starhub connection uses WEP and my Singnet uses WPA. It’s supposed to make a slight difference, but I don’t expect it will matter much because speedtest.net results tend to vary within a few minutes of each other (I’ve run tests consecutively and seen very different results – generally if it doesn’t make sense [like upload speeds coming very close to download speeds], I’ll run them again and take the higher one). It’s not an exact science but I hope the data will be useful anyway.
Note: Starhub product literature mentions that a “speedboost” is provided when visiting certain partner sites, which accounts for exceptional download speeds in some cases, like the Asian results. I’ve left these in because I chose global test servers at random, so all’s fair. Unless of course Starhub partnered these sites specifically to optimize their speedtest.net results. Just saying.
11:30am to 12 noon
(Singnet results before Starhub)
Singnet vs Starhub: 9251kbps vs 7809kbps
Morning European Winner: Singnet
Singnet vs Starhub: 6826 vs 16490
Morning Asian Winner: Starhub
Singnet vs Starhub: 9954 vs 10900
Morning American Winner: Starhub (just barely, and only because of a “speedboost”)
Morning Local Winner: Starhub
Overall Morning Winner: Starhub
I’m switching to Numion.com for this test. It loads a mix of popular websites in a bunch of inline browser frames to measure your “real-world” browsing speed, as well as a regular download/upload test to gauge maximum speeds.
My custom settings for this are: English language sites only, from Singapore, the UK, and the USA. It runs for 30 seconds instead of the default 15.
|Surfspeed inside Singapore:||253.82||(31.72)||51%|
|Surfspeed average (worldwide):||152.05||(19.00)||100%|
|Surfspeed outside Singapore:||145.01||(18.12)||100%|
|Surfspeed inside Singapore:||627.27||(78.40)||100%|
|Surfspeed average (worldwide):||237.16||(29.64)||100%|
|Surfspeed outside Singapore:||225.17||(28.14)||100%|
Starhub wins at surfing for now, but it’ll be interesting to see what results are like during the peak evening hours. When it comes to overall maximum throughput and latency though, Singnet actually takes this one. That might be of some value to gamers and people using P2P systems.
Edit: I just ran the Numion test again for Singnet and this time it managed 987.87kbps for “Surfspeed inside Singapore”, against the previous 253.82kbps, and beating Starhub’s 627.27kbps. The latency for Singnet jumped to 1000ms though. Once again, it just shows how you can’t rely on these tests too much.
I ran another batch of speedtest.net tests, but decided that their tests are flawed and fluctuate too much. At some points, running the same country’s test will route your data through another ISP/host in their network, leading to dissimilar results (I did not take note of these hosts’ names, so could not be sure that I was comparing Singnet & Starhub on equal ground. Numion’s methodology makes more sense.
9pm Numion Test
|Surfspeed inside Singapore:||757.11||(94.63)||39%|
|Surfspeed average (worldwide):||143.31||(17.91)||57%|
|Surfspeed outside Singapore:||136.94||(17.11)||61%|
|Surfspeed inside Singapore:||1,933.16||(241.64)||100%|
|Surfspeed average (worldwide):||251.39||(31.42)||100%|
|Surfspeed outside Singapore:||217.54||(27.19)||97%|
Singnet wins again on maximum speed/latency, but loses out in surfing speed. I’m not sure if this has something to do with my Starhub going through a dedicated 3com OfficeConnect router while Singnet goes through the provided 2WIRE brand integrated ADSL modem & wireless router.
Anecdotally, I’m seeing better perceived performance with Singnet, with YouTube and GameTrailers.com videos starting to play almost instantly.
Downloading the demo for Aquaria from Ambrosia SW’s FTP server. Singnet gives me approx 50kb/s, while Starhub is fluctuating between 15-30kb/s, with more time spent between 15-20kb/s.
I wanted to do a test at 11pm, but couldn’t spare the time. Anyhow 1am isn’t too bad, most people have gone to bed, leaving it the domain of students, gamers, and the unemployed (power users).
|Surfspeed inside Singapore:||896.36||(112.04)||46%|
|Surfspeed average (worldwide):||130.76||(16.34)||52%|
|Surfspeed outside Singapore:||115.63||(14.45)||51%|
|Surfspeed inside Singapore:||952.79||(119.09)||49%|
|Surfspeed average (worldwide):||166.38||(20.79)||66%|
|Surfspeed outside Singapore:||157.93||(19.74)||70%|
I don’t know what’s up with Singnet’s latency calculation, it’s probably faulty or a sign of the 2WIRE modem’s inability to handle opening 40-something http connections at once. Pinging individual servers reveals a much lower number, like 190ms (google.com) or 214ms (twitter.com).
We’re now at the end of the day, and I’m not sure how to read these numbers but I can say Singnet feels as fast as or faster than Starhub for regular surfing. Starhub’s ping times aren’t so great for gaming (but conventional wisdom has always held that DSL is the better choice), but if you visit a lot of Asian/local sites, their speedboost partnerships may be reason enough to put up with high pings and time-of-day-specific slowdowns. Starhub’s recent upgrade to 768kbps upload speeds (on the 8mbps plan) is also very attractive, and I’d like to see Singnet raising their upload caps in the near future.
Your milage will vary, especially since Starhub’s speeds are dependent on location and neighborhood activity. My area isn’t very densely populated, so the Starhub results don’t look too bad at all against the horror stories I’ve heard. If I could choose to keep either one, I’d take Singnet for being more consistent throughout the day and having better response times, but would feel a little sore about having lower upload speeds. I don’t feel that Starhub’s better surfing speeds on paper make a big difference in real world usage. For example, I just loaded the entire front page of the New York Times in about 3 seconds, minus the ads.