Bufferbloat on BRSK connection

Shaolin Chicken said:
Thank you for the detailed reply @Rainmaker. I tested on Cloudfare on good a good rating for video streaming, Average for gaming and good for video chatting.
I don't do too much online mutiplayer gaming, so will not worry too much.
If I do notice a real problem, I will buy the Flint 2 as recommended.

Changed my mind and ordered a Flint 2 :D, now I don't have to think about bufferbloat again.
Click to expand...

I did the same.
I contacted Zen Internet. If it was possible to get funny looks down a phone then I got them, so I bought a new router. Installing Cake on the router, and five minutes messing about determining the best bandwidth, and my latency is A+ rated now.
 
Last edited:
I didn't even realise this was a thing that could be easily measured... Just done a test & I get a rank "F". Although I think a lot of that will be due to running via a wifi mesh at the moment due to not having installed any infrastructure in my house since I moved in a few weeks ago. I'll have to try connecting something via a cable & see how it changes.
 
I realised I have a VM running on my server I can run the test on - Result of a "C". Not perfect, but a lot better & probably something I'm happy enough with for now. I'm running a Mikrotik Hex as my router on an Openreach connection, so I might have to do some digging to see if there's any way to improve this with my current HW.
 
Interestingly, I just had some trial periods with Zen Internet, and discovered something quite revealing.

I mentioned before in this thread that my latency wandered up to 250ms or so on a Zen 300Mbps service, via Openreach. This prompted me to use Cake, which was very effective and reduced the problem by 90%

However, testing different services from Zen, at 500Mbps, using the same Openreach connection, the problem almost completely disappeared. Just a marginal increase of about 5ms or so.

My first reaction was to blame whatever is in the local cabinet, but on further thought, I suppose it could be more complex than that. But the point remains that on a faster Openreach connection, the problem reduces. Now I'm running without any form of QOS and it's all working just fine. Just as effective as Cake is at 300Mbps.

It's very interesting to me that a small difference in service speed can make such a difference to Bufferbloat.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom