First Look: Real-World Broadband Speed Tests in 2000

ISPs love to throw around big numbers. "Up to 1.5Mbps!" "50 times faster than dial-up!" "Blazing-fast downloads!" But what do you actually get when you plug in a DSL modem or cable modem and start using the internet like a normal person?

We spent two weeks testing three connections in the Washington, D.C. metro area: a Bell Atlantic Infospeed DSL line (advertised at 640kbps), a Road Runner cable connection (advertised at "up to 1.5Mbps"), and a baseline 56k dial-up connection through a V.90 modem on a clean phone line. Same computer, same test files, same times of day.

Here's what we found.

The Test Setup

Our test machine was a Pentium III 550MHz with 128MB of RAM running Windows 98 SE — a pretty typical home PC in mid-2000. For the DSL and cable tests, we connected via Ethernet. The dial-up test used an internal 56k modem.

We measured three things:

1. Sustained download speed — downloading large files from well-connected servers
2. Web page load times — loading real websites as a user would experience them
3. Peak vs. off-peak performance — testing at 10 AM, 3 PM, and 9 PM on weekdays

All speeds were measured using actual file transfer times with a stopwatch and a byte counter. No synthetic benchmarks — we wanted to know what real usage feels like.

Raw Download Speeds

We downloaded a 10MB test file from a fast FTP server at three different times of day. Here are the average results:

| Connection | 10 AM | 3 PM | 9 PM | Average | |-----------|-------|------|------|---------| | 56k Dial-Up | 44kbps | 42kbps | 38kbps | 41kbps | | Bell Atlantic DSL (640k) | 585kbps | 571kbps | 578kbps | 578kbps | | Road Runner Cable | 1,340kbps | 1,180kbps | 890kbps | 1,137kbps |

A few things jump out.

First, the dial-up connection never actually hit 56k. This surprises nobody who's used a modem — FCC regulations limit upstream modem speeds on the phone network, and real-world conditions rarely allow for theoretical maximums. Getting 41-44kbps is pretty typical for a good 56k connection.

Second, the DSL line was remarkably consistent. The speed barely varied between morning, afternoon, and evening — hovering around 578kbps regardless of time. That's about 90% of the advertised 640kbps, which is better than we expected. The dedicated nature of the DSL connection means your neighbor's usage doesn't affect your speed.

Third, the cable connection was fast but variable. At 10 AM on a weekday, Road Runner screamed along at 1.34Mbps — close to its advertised maximum. By 9 PM, when presumably most of the neighborhood was online, it dropped to 890kbps. That's still fast — much faster than the DSL line — but it's a 33% speed reduction from the morning peak.

Real-World Web Browsing

Speed tests are useful, but they don't capture what it feels like to browse the web. We loaded 10 popular websites — CNN.com, Yahoo, Amazon, eBay, ESPN, CNET, Weather.com, Hotmail, MapQuest, and Slashdot — and timed how long each took to fully render.

Average page load times across all 10 sites:

| Connection | Average Load Time | |-----------|------------------| | 56k Dial-Up | 22.4 seconds | | DSL (640k) | 3.1 seconds | | Cable (Road Runner) | 1.8 seconds |

The difference between dial-up and either broadband option is enormous. But the difference between DSL and cable? Less dramatic than you might think. The gap narrows on web browsing because page loads are limited by factors beyond raw bandwidth — DNS lookups, server response times, the number of elements on a page. Once you have "enough" bandwidth, the bottleneck shifts elsewhere.

Both DSL and cable make the web feel responsive and snappy. Dial-up makes it feel like wading through mud.

The MP3 Test

Because this is the year 2000 and everyone wants to know: how fast can you download music?

We downloaded a 4MB MP3 file (about 4 minutes of music at 128kbps encoding):

| Connection | Download Time | |-----------|--------------| | 56k Dial-Up | 13 minutes, 20 seconds | | DSL (640k) | 56 seconds | | Cable (Road Runner) | 24 seconds |

On dial-up, downloading a single song takes longer than listening to it. On DSL, it's under a minute. On cable, it's practically instantaneous. If you're a Napster user (and let's be honest, a lot of people are), broadband transforms the experience from tedious to addictive.

Software Updates and Large Downloads

We also grabbed a 30MB file — roughly the size of a browser update or small game demo:

| Connection | Download Time | |-----------|--------------| | 56k Dial-Up | 1 hour, 40 minutes | | DSL (640k) | 6 minutes, 50 seconds | | Cable (Road Runner) | 3 minutes, 30 seconds |

At dial-up speeds, a 30MB download is a commitment. You start it, go do something else, and hope nobody picks up the phone. On broadband, it's a bathroom break.

Upload Speed Matters Too

We uploaded a 2MB file to an FTP server:

| Connection | Upload Time | |-----------|------------| | 56k Dial-Up | 6 minutes, 15 seconds | | DSL (640k, 90kbps up) | 2 minutes, 55 seconds | | Cable (Road Runner, 256kbps up) | 1 minute, 3 seconds |

Upload speeds are the neglected stepchild of broadband. Both DSL and cable connections are heavily asymmetric — fast downloads, slow uploads. The DSL line in our test offered only about 90kbps upstream, while Road Runner delivered around 256kbps. Neither is fast by any stretch, but both beat dial-up for emailing attachments or uploading photos.

Latency and Online Gaming

For gamers, raw bandwidth matters less than latency — the round-trip time for a data packet to travel from your PC to a server and back. We pinged several game servers and a general-purpose server (yahoo.com):

| Connection | Ping to Yahoo | Ping to Game Server | |-----------|--------------|-------------------| | 56k Dial-Up | 180-250ms | 200-350ms | | DSL | 25-40ms | 30-60ms | | Cable | 15-30ms | 20-50ms |

Both broadband connections deliver dramatically lower latency than dial-up. For Quake III Arena or Unreal Tournament players, this is the difference between competitive gameplay and frustrated rage-quitting. The cable connection had a slight edge in latency, but both are in the "very playable" range.

The Verdict

Cable wins on raw speed. DSL wins on consistency. Both absolutely destroy dial-up in every metric that matters.

If you have access to both and price is similar, cable internet will give you faster downloads and snappier browsing most of the time. The evening slowdowns are real but manageable — even at its worst, cable was still 50% faster than DSL in our tests.

If you value predictable, steady performance and don't mind slightly lower peak speeds, DSL is a solid choice. The rock-steady speeds are genuinely nice, and the dedicated connection means you'll never wonder why things are slow at 9 PM.

And if you're still on dial-up? Either option is a life-changing upgrade. Stop debating DSL vs. cable and just get whichever one is available first. The difference between 41kbps and 578kbps is infinitely more meaningful than the difference between 578kbps and 1,137kbps.

Your internet experience is about to get a whole lot better.