{{
`${userDetail.user.displayName}${userDetail.user.cachedBungieGlobalDisplayNameCode ? `#${standardizeBungieGlobalCode(userDetail.user.cachedBungieGlobalDisplayNameCode)}` : ""}`}}
{{getUserClans()}}
由dtdionne編輯：
11/21/2015 1:59:37 AM
1
Hrm, technically, layer 1 and the first half of layer 2 (media access control portion of the data link layer) inner-workings are transparent to the upper layers. There must be more going on, or to this.
I just ran a test...
Background - my ps4 connects to a cisco 4948ee, which connects to a cisco 6807 VSS cluster, which connects to a cisco asa 5585x cluster, which converges (connects) 3 different wans, 1) cisco asr 9022 vss cluster dwdm otn gigapop 2) cable modem resi 3) adtran atlas frame over atm telco
Nothing I did had any effect remotely close to that described, as I expected...but here's what I tried. I decreased mtu size, in 100B increments, on the following interfaces...
1)ASA Inside to 6807
2)ASA Outside (all 3 wans)
3)6807 interface to 4948 and visa versa
4)4948 interface to ps4
So, I presume that this "exploit" is implemented on the ps4 itself...something I didn't and won't try because of the ps4's network abstraction. This leads me to believe that if this is true, it has way more to do with how the ps4 wraps frames than it does layer 1's maximum transport unit.
Regardless, this is still a very intriguing topic, I'd like to know more...
As an asside, Im in the process of writing a crucible specific wireshark dissector. Well, I already wrote one, but it's pretty crude...so I'm writing a newer, and hopefully far more elegant one. One that specifically dissects cheating. If anyone's interested, I'm down for discussion...if not, that's cool too. The ultimate goal is to document the "anotomy of crucible cheats" in a video or something
English
-
The exploit required changing the MTU on the actual PS4 or local router for an Xbox, the issue was not the networking path but how the game handled those packets. There have been a lot of patches on both the consoles and the game since then. The main issue with your test come at you are playing different versions of the game. I can't run a Service pack 1 exploit on Service pack 2 if that new version modified the parameters being exploited. Since August, there has been a lot of changes to the network requirements. For example the time out before a player that was lagging is orbited was changed. That then took out the part where you could load into a crucible map to just capture zones, because you would be orbited after a few minutes. Cheaters have since evolved into using Mods, a while back before a player was finally caught they apparently modified the game to have infinite ammo, couldn't be killed, increased the damage of any weapon that allows it to 1 shot kill, unlimited heavy ammo at the start of the match, etc... That guy published his code and a lot of script kiddies kept hammering away to find out what Bungie detection systems were and then adjusted the software to under Bungies' parameters. People have gone a long way... there is even a mod where you take damage only till you reach 1/10th your health to make people feel like they are missing... One host script went as far as not allowing anyone to aim at the host, meaning your own system would push your cursor away from a box around a target. The only way to grab the data to detect cheating on your end requires you to be host in ever match, that way you get all the packets except the two links going to bungie. The issue comes that most of the exploits require that user to force host, if they are host then you only see your packets and the packets coming from the host. Thus, running wireshark, C&A or any other packet filtering/debugging tool would only be useful to the host. The P2P design is massively flawed. The main issues comes that the penalties are not serious enough to prevent people from continuing to exploit the game and that P2P allows for too much control over the game on the host system. A game like this has to be played on a dedicated server, otherwise there is no integrity.
