You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
Shadowsocks 是非常好用的 代理软件 在,它并不一定是只能用来突破网络审查。
在一些情况中它可以作为轻量、部署简单且功能强大的内网代理。
而在这些环境下并不一定需要强加密,如 rc4-md5 加密方式性能损失非常低。
如在内网环境下可以通过 shadowsocks with rc4-md5 实现简易、高性能的跳板机连接。
这种情况下无需担心协议会被嗅探。
Describe the solution you'd like
希望加回不安全但非常高效的加密方式,或者默认隐藏这些加密方式,然后通过开关来控制这些选项是否显示(标注这些方式是不安全的、容易被嗅探的)
@database64128 Since the decision has already made, I recommend we just lock this session and stop those nonsense talking & arguments. If someone have any other idea, they can "satisfy their own requirements" open a PR like @EpLiar and @kslr , and wait for some Jan Doe to approve it, if lucky enough :)
I will just have my own RC4 supported "unsafe" clone in my local Git repo, since the "community" does not seems happy for the patch.
Activity
kslr commentedon Jan 4, 2021
在大多数情况下 aes 比 rc5-md5 性能更好
chenshaoju commentedon Jan 4, 2021
你仍然可以使用旧版本的客户端和服务端,这一点没有区别。#3041 (comment)
TGSAN commentedon Jan 4, 2021
在测试用的 arm32 架构轻服务器下 aes 的 CPU 占用比 rc4-md5 要更高,目前已回滚至 4.3.3 继续使用。
zonyitoo commentedon Jan 4, 2021
所使用的arm32架构服务器,是否支持AES指令?使用的服务端实现是哪个?
TGSAN commentedon Jan 4, 2021
不支持 aes 硬件加速,服务端实现是 shadowsocks-libev
davendu commentedon Jan 4, 2021
从安全上网角度考虑,使用专用的跨国专线上网更加安全,且快捷(虽然并不方便,就像没有了rc4-md5的shadowsocks一样)。
而且从用户角度来看,是要考虑攻击者的成本的。完全有理由假定在内网环境中,攻击者可以访问/记录所有流量,但并不能解密流量,也没有动机对协议类型做嗅探。
DuckSoft commentedon Jan 4, 2021
本来不想提的: redirect attack on shadowsocks stream ciphers
davendu commentedon Jan 4, 2021
协议可以被解密 !== 攻击者能解密
攻击者也是有自己的困局的,比如流量超过了攻击者能处理的带宽,或者攻击者只(能)关注明文流量。请注意,我关注的是有局限的内网环境,而不是公网环境。
24 remaining items
guoguo-2026 commentedon Jan 4, 2021
但你也忘记回答
fakeboboliu commentedon Jan 4, 2021
嘛,上面一位都说出来“生态圈”了,那生态圈内做事应该统一啊,不把其他 impl 的 rc4 也删一下?
davendu commentedon Jan 4, 2021
@studentmain
我认为你的这条回复典型代表了:不就事论事,阴阳怪气,以及对我和issue中其他讨论者的极度不尊重。
既然您对Fistel这么有研究,不如来向我们传授一下同样是使用了Fistel网络,为什么我们不能在保证密钥长度足够长的前提下应用DES算法。
@SekiBetu
事实上的确有一部分企业的应用收到了影响。
@kslr
这是issue中的讨论,不是需求单。
我觉得这个讨论串真的很有意思,一直在用户明确认为可以接受不安全的情况下讨论这个做法是否安全,甚至连加入讨论都需要一个“熟悉这个圈子”的资格,中文互联网的圈子精神被发挥的淋漓尽致。
另外你们大可以明确的把“保留这些算法会增加我们的维护量所以我们决定不维护了”说出来,虽然我们都知道 #2757 本身也是工作量的一种体现。
@database64128 Since the decision has already made, I recommend we just lock this session and stop those nonsense talking & arguments. If someone have any other idea, they can "satisfy their own requirements" open a PR like @EpLiar and @kslr , and wait for some Jan Doe to approve it, if lucky enough :)
I will just have my own RC4 supported "unsafe" clone in my local Git repo, since the "community" does not seems happy for the patch.
guoguo-2026 commentedon Jan 4, 2021
您可以就行杠精,可以不尊重我的观点和看法,不如回复下之前被你无礼搪塞@davendu 他的留言
zonyitoo commentedon Jan 4, 2021
Discussion here are kind of pointless. My friends, be rational!
DuckSoft commentedon Jan 4, 2021
Just lock the thread. This is not discussion. @madeye
SekiBetu commentedon Jan 4, 2021
原issue认为性能会有损失,那么,想办法支持AES不就好了,既保证了安全也保证了性能,而不是把精力花在支持不安全的方法上
davendu commentedon Jan 4, 2021
@SekiBetu 我觉得这也是一种方案……但那就要找SoC厂商流片了
davendu commentedon Jan 4, 2021
请试图在网页中搜索您的答案
SekiBetu commentedon Jan 4, 2021
那我只能是说,相信未来会有更多设备会支持AES了,这应该是个趋势吧