Methodology problem with M-Lab testing

TechsysTechsys Member Posts: 5
First Comment Photogenic
This is from an email i sent to M-Labs Support:

"I live and work in the Houston, Texas, area. I strongly object to M-Labs results which appear to favor Comcast but fails to distinguish between AT&T customers using Fiber connectivity versus AT&T customers who still use DSL or possibly even T1 spans. That’s like comparing LAN speeds to dial-up! Combining every type of connectivity offered by AT&T is inherently unfair!

AT&T has been actively deploying Fiber to virtually every part of Houston for at least the last 5 years and thousands of homes have become subscribers who, like me, enjoy more than 400 Mbps downloads on a regular basis but in BOTH directions. It is NOT possible for Comcast to do this over coax!

I also don’t understand why AT&T Mobility is in the list because that is cellular. If you’re going to do that, then why isn’t Verizon, T-Mobile, Sprint, Consumer Cellular, Straight Talk, or any other cellular Internet provider listed?"

VioletChepil

Comments

  • GidsterGidster London, UKMember Posts: 224
    100 Likes 100 Comments 25 Awesomes 25 Agrees
    ✭✭✭
    I assume you're referring to our Scoreboards, rather than any ranking M-Lab (our supplier) is giving. Our initial scoreboards are pretty simple and we appreciate any feedback you have to improve these in future.

    You're correct in pointing out that they don't differentiate providers with very different infrastructure. An ISP that only has access to ADSL will never be able to achieve speeds anywhere close to networks that have access to coax cable or fibre. It's like in the UK, where Virgin's coax network can offer 1gbit/s to some customers but the best those ISPs reliant on BT Openreach's ADSL and Fibre is often 70mbit/s. But perhaps companies like Virgin should be rewarded for the investment they've made in their infrastructure.

    We do also ranking by customer satisfaction, and often those rated highest here are not those offering the fastest connections.

    You're also correct that we are listing mobile networks next to broadband networks and we realise that this too is comparing apples to oranges - even more so as performance is effected so much by environmental conditions and contention.

    Head of Product at Fing
    VioletChepil
  • TechsysTechsys Member Posts: 5
    First Comment Photogenic
    Actually, the monthly email with subject "Internet Performance Report" yet again reported low "ratings" for AT&T and has been relatively unchanged for many months now. This prompted me to investigate the source, Measurement Labs. I found their graphical query utility and entered Houston, Texas, at https://viz.measurementlab.net/location/naustxhouston?isps=AS13367x_AS10774x_AS20057 which reflects surprisingly low transfer rates overall but also presents data in a way that I find very questionable because the results are egregiously skewed against AT&T. That graph prompted my complaint. Having enjoyed more than 5 years of AT&T Fiber service, I have experienced exactly one outage due to a natural event. Too, I regularly achieve higher that 800 Mbps transfers within the Houston metropolitan area. I do not feel the data presented by M-Labs is reflective of actual results I have personally observed and it should be investigated.
    VioletChepil
Sign In or Register to comment.