|
发表于 2011-12-15 18:41:53
|
显示全部楼层
我贴一个测试结果:
服务器是在沈阳联通,带宽是100M共享的,测试客户端是上海电信,10M光纤
网络ping的情况:
Ping statistics for 218.25.x.x:
Packets: Sent = 20, Received = 15, Lost = 5 (25% loss),
Approximate round trip times in milli-seconds:
Minimum = 146ms, Maximum = 148ms, Average = 147ms
没开启加速:
C:\Documents and Settings\Administrator>wget -O a.txt http://218.25.x.x:8000/lastb.txt.gz
--2011-12-15 18:53:45-- http://218.25.x.x:8000/lastb.txt.gz
Connecting to 218.25.x.x:8000... connected.
HTTP request sent, awaiting response... 200 OK
Length: 8337954 (8.0M) [application/octet-stream]
Saving to: `a.txt'
100%[======================================>] 8,337,954 16.0K/s in 83s
2011-12-15 18:55:09 (98.0 KB/s) - `a.txt' saved [8337954/8337954]
加速后:
C:\Documents and Settings\Administrator>wget -O a.txt http://218.25.x.x:8000/lastb.txt.gz
--2011-12-15 18:47:22-- http://218.25.x.x:8000/lastb.txt.gz
Connecting to 218.25.x.x:8000... connected.
HTTP request sent, awaiting response... 200 OK
Length: 8337954 (8.0M) [application/octet-stream]
Saving to: `a.txt'
100%[======================================>] 8,337,954 432K/s in 19s
2011-12-15 18:47:42 (424 KB/s) - `a.txt' saved [8337954/8337954]
[ 本帖最后由 aru 于 2011-12-15 18:54 编辑 ]是否有安全隐患我不清楚,这是一个内核模块,也就是说,如果他们有恶意代码,对方是可以搞定你的服务器的。
200M内存很正常啊,这个是对于大流量服务器用的。200M对于100Mbps来说也就缓存20秒的数据而已
原帖由 windywinter 于 2011-12-15 18:40 发表
他们的tcp栈跟hybla相比怎么样?
提醒尽量不要乱用来历不明的tcp协议栈,如果协议栈不完全符合rfc标准,则可以制造出安全隐患来。一个协议栈使用200M内存已经很值得怀疑了。 ... |
|