Dude, have you even read what I wrote? Clearly not.
1.) I fixed it. With the help of that article.
I have read what you wrote but I can't say that you've read what I wrote nor the article you used. Read both again and try to comprehend what both say. From the information from you and that article I can't tell that it was a Lion/OS X upgrade problem and you actually fixed it. It might be the case but I can't tell for sure.
2.) I haven't used any limiting speed tests, I have used the one provided by my ISP. They use it as a support utility and it is clearly not limited. In fact, if you call them to complain about speed issues, first thing they do is point you to that very speed test. They wouldn't do that if they had limited it, would they? Think.
Yes please think before you do something. You make a very common mistake beginners that have no idea what they are doing make: they trust tests from 3rd parties like their ISP. If there is 1 kind of test you should never ever trust are speedtests, especially when they are from your ISP! What do you think what happens when that test shows the reality, that people are not able to get the full speed of what the ISP says they offer? It is in the ISP's best interest to not display the actual speed but a far more positive one. If you do multiple speedtests from various 3rd parties and tools you'll see the difference. So yes, please think about what you are doing and try to comprehend how things work.
But, if you have actually read my reply you'd have known I was talking about the technical part, not about the marketing part. That means I was talking about technical issues where some servers are slower than others. One of those things are the network links: their speed AND their LOAD (!!!!). If lots of people are connecting to that server and it can't keep up you will get slower speeds. It is like a bicycle: when you ride alone you can go very fast. Carry somebody and you won't be riding as fast. Same thing for other kind of vehicles.
If you would have actually read my reply you'd also have known that I was talking about limiting your connections bandwidth and NOT the speedtest. This is common networking (see, it helps if you actually know what you are doing instead of copy-pasting stuff). ISP's offer different subscriptions with different bandwidths. When using something like fibre connections they run standard fibre cables that have a bandwidth of 10 Gbps. They won't give you that though. Instead they will limit the bandwidth to whatever is in your contract. Sometimes they screw up (I've seen it happen quite often) and people get stuck on a lower speed. If something is limited to a fixed speed of say, 50 Mbps it should ring some bells. In case of problems it is less likely that you'll see such a limitation, it is more likely that it will fluctuate. To give you an example: the Dutch ISP KPN offers fibre connections (ftth). You get internet (very fast ones), tv/radio and telephony. What they do is put separate VLANs on your connection: 1 for tv/radio, 1 for telephony and 1 for the standard internet connection. They do this because they can limit the bandwidth per VLAN. This way your internetconnection, telephony and radio/tv do not interfere. This means you can watch an HD movie, download like there is no tomorrow and make normal phone calls. They also offer these services on their DSL connections. They can't do it there and it causes problems when you try to do all these things. One of the reasons why most people on DSL will be limited to SD.
3.) Most importantly: When I tried the same speed test utility from my ISP using Windows 7 instead of Mac OS X (same day time, two minutes after the test with Mac OS) I got the full 100 mbit. Two minutes later I try again with Mac OS, I get 50 mbit. This had a clear pattern and was reproducible.
It is very common that these kind of tests have fluctuating results. You need to take quite a lot of tests and from different parties to have some kind of certainty.
4.) I tweaked the TCP parameters as described in that article and the speed test *immediately* went up to 100 mbit. A clear indicator that whatever was set before, was less than optimal.
Or the ISP fixed something or the load in the server was less or the load on the network between you and the server was less or... There are way too many variables that can explain why it is now fixed. You should have used various tests, not just 1. By using more than 1 you minimise this risk of problems.
Also, I never said this would be a isolated Lion problem. All I said was that this happened after I upgraded from 10.6 to 10.7. I did not have this problem with 10.6.
Just wanted to point out that this isn't a 10.7 problem but simply an OS X problem. 10.6 has the same kind of problem. The fact that you used 1 test is the big problem. If something is wrong with that test it will affect the results. If it spits out wrong results you come to wrong conclusions.
I never had problems on my local network. I am using iSCSI storage all the time with full 1 gbit speed with no problems. The problem I described only affected connections that traversed a router, e.g. my internet connection.
The more reason to doubt that you actually fixed it by using those settings.
