ISPs Are Mostly Delivering The Speeds They Advertise, Just Not Consistently

realvadsWhat does it mean when a cable company advertises “blazing fast Internet” or download speeds “up to 15 Mbps”? Does that mean all the time for everyone, or just an average? And how far from those “up to” speeds can an Internet service provider be before they have some explaining to do?

The FCC has released its latest Measuring Broadband America report for 2014, which looks at the actual speeds that American consumers are getting from the nation’s largest ISPs and how those stack up against what those ISPs are promising.

The good news: For the first time in the four years that the FCC has been doing this report, ISPs as a whole are delivering the speeds they promise… on average.

As the chart at the top shows, some ISPs are doing better than others. For example, Verizon FiOS, Comcast, and Cablevision are each delivering sustained uploads and download speeds that meet or exceed the advertised rates. Others, mostly DSL providers are still not able to deliver the speeds they promise. But even the worst provider in the study, Windstream, is still coming in above 80% of its advertised speeds.

But the FCC’s first look at the consistency of these speeds shows some cracks in the system.

This chart can be a little confusing. The standard for consistency set by the FCC for this metric was whether 80% of an ISP’s customers get the full advertised speed 80% of the time. So while Cablevision, Comcast, Cox, Mediacom, FiOS, and ViaSat/Exede all delivered sustained speeds that are above the advertised levels, only Cablevision and FiOS are consistently delivering the promised speeds. Meanwhile, many of the DSL providers — Frontier, Verizon, Windstream — are far from consistent in delivering advertised speeds.

FCC Chair Tom Wheeler has directed the commission to reach out to ISPs that are providing slower than advertised and/or highly inconsistent speeds to find out what these companies are doing to resolve these discrepancies.

“Consumers deserve to get what they pay for,” said Wheeler in a statement. “While it’s encouraging to see that in the past these reports have encouraged providers to improve their services, I’m concerned that some providers are failing to deliver consistent speeds to consumers that are commensurate to their advertised speeds.”

With regard to the hot-button topic of interconnection — the points at which commercial bandwidth providers dump off their data to ISPs to carry the last mile to consumers — and the congestion of interconnection points, the FCC didn’t include any data on that.

The report states that it has historically not counted statistics for data over degraded interconnection points because that had usually just been a correctable network error as opposed to something systemic.

However recent spats between ISPs like Verizon, Comcast, and AT&T and Netflix’s bandwidth providers have affected a large number of people and the FCC recognizes that “consumers accessing services and content over the affected paths would likely see a significant degradation in their service.”

And so the report does not include measurements from degraded interconnection points, but the FCC has collected test results from impacted servers and is releasing this information for others to pick apart.

Chairman Wheeler recently directed his staff to investigate this type of congestion and has obtained details on the deals Netflix made with Comcast and Verizon to pay for better access to their networks.

Read Comments2

Edit Your Comment

  1. Mokona512 says:

    These measurements do not take into account time, it means nothing if on average, the ISP offers 80% of advertised speeds, if the average is made up of part of the day it is fast because everyone is at work or a sleep, and insanely slow during the times people actually want to use their connection.

    Averages often give false representations, and thus speeds must be measured at multiple times per day in order to graph the speed over time.

    A friend of mine has time warner and is advertised as getting 30mbit/s download and gets in the upper 20’s late over night, but in the afternoon and evening, he is lucky to get even 5mbit/s and during those times, the pings are too high to do fast paced multiplayer games. there is a lot more to internet connections than just throughput.

    Overall the issue is not whether they can offer their advertised speeds, but are they able to offer those speeds throughput the day, and not drop you to a tiny fraction of your speeds in the evening in addition to a massive increase in ping times.

    For my fios service, I consistently get 95-100% of my throughput based on speed tests, but serivces like netflix are too slow to get HD, and services like youtube become unusable, I have to use a firefox extension to download the videos and watch them with VLC because the streaming on youtube is to inconsistent in the evening. (one ,moment it will stream fine, the next, it will not load any new data for 2+ minutes. (if I use my phones hotspot with the PC, that issue goes away)

    The test should look into which areas are they providing the advertised speeds. (many speed test servers are hosted by the ISP’s or other companies using their network)

  2. mrkake says:

    And here I am in Boston, one of the few places where I actually get a choice in broadband (between Comcast and RCN). Currently on RCN and speeds are consistently always 100% of the advertised speed! Competition rules!!!!