Ruisu BBR, magic modified BBR, effect comparison, which one is better, Ruisu or BBR?

Installation tutorial

bbr plus, bbr original, bbr magic modification and sharp speed four-in-one script

bbr plus, bbr original, bbr magic modification and sharp speed four-in-one script

Basic situation of testing

Recommend the most cost-effective VPS for tile movers: 1 core/1G/20G hard drive/1T traffic/1Gbps/1IP/multi-machine room switching/CN2 GT line--click to buy

Server: Vultr 1 core 1G memory 1G bandwidth Address: Frankfurt, Germany IP:140.82.37.210 Some merchant packages: (Sign up to get 5 US dollars, complete tasks to get 3 US dollars)

Memory	CPU	SSD	Flow	Price	Buy
1G	1 core	25G	1T	$5/month	Link
2G	1 core	55G	2T	$10/month	Link

The local network is: Henan China Unicom 100M downlink and 20M uplink (both uplink and downlink have burst bandwidth) Route tracking shows that the specific path is Henan Unicom-->Beijing Unicom-->Los Angeles Unicom-->Romania Bucharest-->Germany Frankfurt; there is a slight detour in Europe

No accelerated testing

1:100 PING test delays and packet loss

Through 1,000 ping tests of this machine, we can see that the average delay is 295ms and there is no packet loss. This surprised me. The physical distance is so far and there is no packet loss. It is simply magical!

2: Basic information of HTML5 Speedtest test

The measured download speed gradually increased from 2Mbps to a maximum of 5.88mbps, gradually accelerating, and the acceleration speed was average. The upload speed is basically 18Mbps. It seems that the return trip is congested, but the outbound speed is acceptable! Test results:

Test environment	Download speed	Upload speed	Average latency	Delay jitter
Nude connection	5.88	18.06	380.27	58.71

Install the sharp effect

1:100 ping test

The ping value is average, and the number of packet losses is similar to that of naked connections

2: Basic information of HTML5 Speedtest test

After turning on Ruisu, the download speed starts directly at 7Mbps, the upper limit reaches 50.11Mbps bandwidth, the upload bandwidth is basically full, the average delay is 387.68ms, and the delay jitter is 44.3ms

Test environment	Download speed	Upload speed	Average latency	Delay jitter
Rapid speed	50.11	20	387.68	44.3

BBR test results

1:100 ping test

The PING value and packet loss are similar to the results of the previous test, and there is not much change.

2: Basic information of HTML5 Speedtest test

It can be clearly seen from the figure that in terms of download speed, both the starting speed and the upper limit of download speed are far behind Ruisu.

Test environment	Download speed	Upload speed	Average latency	Delay jitter
BBR	12.72	17.35	390.83	37.76

BBR magic modification effect

1:100 ping test

The PING value and packet loss are similar to the results of the previous test, and there is not much change.

2: Basic information of HTML5 Speedtest test

There were rumors on the Internet that the effect of the BBR magic modified version is amazing, but judging from the actual speed measurement effect, it is only a limited improvement compared to the original version of BBR.

Test environment	Download speed	Upload speed	Average latency	Delay jitter
BBR magic modified version	28.38	17.42	409.33	224.47

BBR violent magic modified version effect

1:100 ping test

The PING value and packet loss are similar to the results of the previous test, and there is not much change.

2: Basic information of HTML5 Speedtest test

BBR violent magic modification, it seems, is even worse than BBR magic modification, but there is still a certain gap in relative speed.

Test environment	Download speed	Upload speed	Average latency	Delay jitter
BBR Violent Magic Revised Version	22.07	14.13	454.46	559.79

Data comparison

test environment download speed upload speed Delay fluctuation Video speed peak Video speed average Naked Rapid speed BBR BBR magic modification BBR violent magic modification

5.88	18.06	380.27	58.71	2000	2000
50.11	20	387.68	44.3	15000	10000
12.72	17.35	390.83	37.76	7000	3000
28.38	17.42	409.33	224.47	6805	2500
22.07	14.13	454.46	559.79	10000+	3200

The PING value will not be shown for now. After several solutions, the change is not big. The particularly large fluctuations in the above table are caused by packet loss. From the above table, you can clearly see that Ruisu crushed BBR, BBR magic modification, and BBR violent magic modification. BBR is an open source TCP acceleration module developed by Google. BBR comes with Google's halo, but in actual use, it is not as good as the domestic Ruisu. Of course, Ruisu is not without its shortcomings, because the Ruisu project has been suspended. Therefore, Ruisu no longer supports higher versions of the kernel. For this test, centos6 was installed on my vps.

Conclusion

1. If the kernel supports it, use Ruisu. If the kernel does not support it, use the BBR violently modified version! 2. No software can save junk lines. Pick a good line first~

Recommended site search: virtual host evaluation network, Hong Kong host, high defense server, Hong Kong vps host rental, Shandong website registration, pw domain name, cloud server hosting, domain name information, Hong Kong high defense server, game high defense server rental, < /p>