One of the most interesting statistics coming out of the latest OpenVault Broadband Insight Report for the second Quarter of 2024 is that the percentage of customers using a terabyte of data each month continues to grow. OpenVault predicts that more than 20% of households will be surpassing a terabyte of data soon.
The following table shows the growth in what OpenVault calls power users – those that use more than 1 terabyte of data per month. The 2 terabyte customers are also included in the first column numbers. There was a one time burst of new terabyte users during the pandemic like there was with overall broadband usage, but the growth has continued since then.
One issue with so many terabyte users is there must be a lot of homes paying extra for broadband because of data caps. That is where an ISP charges more for usage over some predetermined amount. Most landline ISPs now offer unlimited data usage – but not all. Comcast has a 1.2 terabyte monthly data cap that it only enforces in some markets. Cox has a monthly data cap of 1.28 terabytes. Mediacom has data caps that range from 350 gigabytes to 1 terabyte for its various products. Sparklight has a 700 gigabyte data cap on some products but unlimited on more expensive ones.
I recall in the early days of broadband that the vast majority of ISPs worried about large broadband users. This was largely due to the technology in use at the time. For example, I know many early fiber ISPs had deployed BPON that delivered 622 Mbps download and 155 Mbps upload to a neighborhood node of around 322 homes. A few large data users in the same node could cause noticeable problems for their neighbors. Cable companies that deployed DOCSIS 3.0 technology had even more problems since they often had 200 or more customers sharing the same node. Since then, most cable companies have split their nodes and reduced the number of homes sharing the same data connection.
I remember working with a number of ISPs that wanted to know what to do about heavy broadband users. I recall clients who had customers in residential neighborhoods who operated music file-sharing servers, and in one case, a pornography server. ISPs worried about people who operated early successful retail servers. I recall a customer who generated a huge amount of traffic to support a chat line service – something the FCC eventually frowned on. More recently, one of my clients had an unusually large number of customers operating bitcoin servers since the ISP operates in an area with cheap hydropower electricity.
In almost every case, the ISPs decided to invite heavy users to upgrade to a business service, with the extra payments helping to fund setting up an extra node. In a few cases, the ISPs ended up disconnecting customers they deemed were harming the network. This was always a gray legal area, and every ISP who did this struggled with the decision.
This is less of an issue today with more robust networks. Fiber ISPs now usually have the option to serve a few customers in each node with active Ethernet to segregate them from other customers. I’ve always wondered how the big cable companies handle this – it’s not something I’ve ever heard them discuss.
OpenVault says there is now a growing number of customers who use more than 5 terabytes of data per month. They didn’t cite the percentage, but anybody using that much broadband is tying up a lot of node resources. And I’m sure there are some customers using far more than 5 terabytes. I still think it’s a valid question to consider if somebody who regularly using that that much data is really operating a business and should pay business rates.
I can answer that last question for me personally. I have 3 children, me, my wife, and regular kids that come by the house to see my kids. My home data usage last month according to my plume account was 7.2 Terabytes. We conduct zero business out of our home.
I’ll even post my stats
Data downloaded with this location
225.47 GB
in last 24 hours
Usage rating: HIGH
1.77 TB
in last 7 days
Usage rating: HIGH
7.28 TB
in last 30 days
Usage rating: HIGH
Graph shows a % comparison usage between other times of the total usage, to determine popular times and days. This location’s WAN saturation can be seen on the Health page, and more detailed WAN real usage graphs are on QoE page.
24 hr usage
18.16%
40.95 GB
2:00 PM
16.12%
36.34 GB
6:00 PM
17.76%
40.04 GB
10:00 PM
16.23%
36.59 GB
2:00 AM
16.29%
36.72 GB
6:00 AM
15.45%
34.83 GB
10:00 AM
7 days usage
11.36%
205.39 GB
10/3/2024 Thu
16.08%
290.69 GB
10/4/2024 Fri
15.18%
274.39 GB
10/5/2024 Sat
14.24%
257.35 GB
10/6/2024 Sun
14.21%
256.88 GB
10/7/2024 Mon
15.88%
287.04 GB
10/8/2024 Tue
13.05%
235.81 GB
10/9/2024 Wed
Less
The ‘problem’ here is that monthly use has no bearing at all on the cost to deliver the service, it’s peak primetime use.
ie, ISPs buy unmetered links at a given ‘momentary throughput’, the only use that makes sense to take into account is that peak use during busy times because the higher that is, the larger the uplink connection they have to buy.
If a customer pulled 10Mbps from 7-8pm everyday, they’re the same as 10Mbps 24×7, it’s only that prime time or peak hours use the increases the ISPs required uplink size.
so measuring total delivered data is completely arbitrary.
Now, metering services can pummel a customer into more conservative use which could translate into lower peak hours usage, but what a round about way to do this.