Comcast Gigabit performance evaluation Check the speed and reliability of your connection with this review.
Comcast Gigabit performance evaluation Check the speed and reliability of your connection with this review.
They expected a smooth experience with their current setup, but it didn’t work out. I’m hoping the fiber connection will be better for me.
You need to grasp how things function before expecting changes. That’s similar to demanding why your CPU isn’t 10 GHz right out of the box since 3 GHz models were released 15 years ago. In a proportional sense, it should just work fine! Expecting it never to be an issue is overblown. There’s minimal impact on users unless you’re handling large data transfers—something most people don’t do. The DOCSIS standard keeps evolving to tackle some of these issues, but eventually physical constraints come into play. For example, with fiber optics, you just need to adjust the equipment to boost bandwidth. Keep in mind that the average global internet speed is around 10 Mbps, so it’s pointless to grumble about 40 Mbps. It’s a waste of time to complain when you can’t justify spending more than 50 Mbps yourself. My biggest complaint about Comcast is their terrible WAN peering. Not your problem, Comcast—your download speeds are probably heavily affected.
Well, you're correct. It makes sense to investigate further. I don't want to be an uninformed buyer, that's my goal. I checked the upper limit for DOCSIS and found this: a 3.0 upstream capacity of 200 Mbps. That’s around 200 Mbps coming in from upstream. Since 2006, the standard is now 12 years old. It also marks the first time gigabit speeds were available. This aligns with the 32 versus 8 channel bonding discussion. Eight channels at 32% equals 25% of the total capacity. That translates to 250 Mbps for a 1 Gbps link. The 70 Mbps figure represents about 7% of the downstream bandwidth, meaning the 8-channel setup isn’t operating at full potential. It seems the modems are designed with 32 channels going down and only 4 upstream to cut costs. That could reflect budget constraints rather than a physical limitation. Unless I'm overlooking something, this explanation fits what I see. I pay for download speeds I choose, and my location doesn’t restrict me from getting up to 800 Mbps down or 800 Mbps up if needed. I’m not constrained by the company’s bandwidth limits here. I could achieve speeds up to 800 Mbps in either direction without being restricted. I’m close to matching upload speeds of a gigabit plan, but only about four times the cost. Also, your own note on global averages—10 Mbps—makes sense when you compare it to the 50 Mbps often cited. In reality, 50 Mbps is closer to the opposite end of the scale. Ultimately, people pay for what they require. I’ll pay for what suits my needs. My priorities are about upload speeds that let me back up PCs in 1–3 hours instead of 10–30, and not needing fast downloads for games in five minutes. I’m okay waiting an hour for a download, and 4K YouTube works fine with a steady 50 Mbps. I could stream three streams at once if desired.
The issue you're facing is that those cable lines have a cap on how much data they can handle. Docsis boosts this by dividing the bandwidth into finer frequency bands per stream, but this only offers a short-term solution. Even if they reach that theoretical limit, the lines would quickly become overwhelmed with just a few users. We're also constrained by the hardware limitations. Docsis 3.0 can handle up to 48mbps per channel, which means for 32x8 channels the modem would be restricted to about 1372/245 Mbps. The challenge remains: you're restricted by how many channels are used at once in these lines. In Docsis 3.1, speeds jumped to 64 Mbps per channel, though upload speeds stayed at 48 Mbps. Once full duplex is unlocked, it will rise to 64/64. For 32/8 channels this pushes us to 2048/245 Mbps, and with full duplex it jumps to 2048/512. The main factors affecting performance are the distance of cable stations and the number of users on a single connection before it switches to fiber. In my area, Comcast typically extends cable to homes that then connect to their fiber network. I struggle to recall the precise capacity per line, but it’s clear there’s a cap on throughput. That’s why they’ve focused on Docsis upgrades to boost capacity—breaking data into more channels across smaller frequency ranges.
The concept seems to focus on how bandwidth is distributed among channels. If total throughput is fixed, adjusting channel allocations doesn't change the overall capacity. It appears to be a strategy to boost download speeds while limiting upload options, which may not align with standard service offerings. This approach could limit performance for users expecting higher upload rates. For more clarity, I’d look into industry reports or ISP documentation on channel allocation practices.
They aren't balanced in any way. They allocate space for uploading and downloading. If it were evenly split, it would restrict the number of download channels they could support, which would slow everyone down overall. This isn’t about limitations but about priorities. Download speeds matter more to most users, so resources are focused there rather than upload. In theory, I agree with their approach. The challenge lies in imagining this scenario for a few hundred users sharing the same bandwidth. Some pay for 150, others for 250, some for 400 or even 1GB—this comes from somewhere, and people seem more concerned about downloads. Upload speeds range from 5 to 35 Mbps, but most don’t rely on them much. For those seeking faster uploads, switching to a business connection could help, though it’s very costly for the benefit received. In short, they’ve likely increased download options because that’s what users value most. Those needing higher upload speeds probably make up less than 2-4% of subscribers, and most of us are among those who don’t.
In simple terms, you could get better performance if you paid enough for a faster upload connection and the ISP agreed to lower their top download speeds for other users nearby. However, ISPs usually don’t do that because they aim to sell faster speeds to more customers, especially since download speeds are prioritized. That doesn’t mean it’s a technical issue—it’s about market choices. I still don’t think it’s a limitation of the technology itself; my need for higher speeds isn’t just about what’s possible, but what’s available in the market.
There are constraints on the hardware regarding upload channels. Even the top modems are limited to 32/8 speeds, which means while 3.1 full duplex could suggest up to 512 Mbps, it still falls short of the 2 GBps download speeds we often see. Cable connections remain asymmetrical, and DSL is similarly affected. Generally, you're left with fiber for truly symmetrical upload and download performance.
Yes, those 32/8 rules apply to consumer modems, correct? Otherwise an ISP couldn't realistically provide gigabit speeds to more than one person in a neighborhood if their network was limited to that bandwidth across the whole street. At 38Mbps per channel in the US, that would translate to about 1,216 Mbps download and 304 Mbps upload. I'm guessing the upload channels use a narrower frequency range since the theoretical cap is 200, not 304, and also accounting for roughly a 10% loss from noise or interference. That brings the practical limit down to around 1.1 gigabits per second download and 180 Mbps upload for your home modem. Does that seem accurate?