PDA

View Full Version : Question on rotating the encryption



Gungadin
03-10-2002, 08:56 PM
I am asking these questions purely for my own information and for the purpose of discussion, and would also like to point out that i am OPPOSED to a windows version of SEQ, for MOST of the reasons that are so commonly stated. Also, i really know nothing about encryption, so i'm hoping so i might learn something. Ok, on to the question.

I have seen many people flamed for proposing a windows version of SEQ, and one of the main arguments is that Verant will start rotating the encryption 1 or two times a week, maybee even once a day. Does this mean that each time they change the encryption , there will be a new version of some client files to be downloaded? (this would be my guess)
If it does not require a new download, then i assume that this means that the client software is build to support a finite quantity of encryptions, and if so surely SEQ could do the same?
IE the faster they change the encryption the faster SEQ developers learn all the available encryptions built into the version.

If it requires a new download every day, how feasable is that for Verant? I mean, if you don't have cable, and you have to do a 30 min download every day before you play then there would probably be a large number of people who would quit. Seeing as how Verant is already getting sensative about reductions in it's player base (ie hiding online population figures) would they really risk alienating their users further?

Your thoughts?

Cryonic
03-10-2002, 09:29 PM
a change of just one file doesn't create a 30 minute download, even on dialup. They didn't make daily changes to the encryption, but made them often enough (about once a week) to foil WinSEQ and SEQ because it would take a day or two for the developers to find the new key, rebuild the libEQ.a file and get it distributed, then VI would change the key again.

The difference is now they would have to change the encryption algorithm because, IIRC, they are now already using a rotating key, everytime you zone the client and server negotiate or exchange a new key. Of course they could always just strengthen the new spawn encryption and that might remove the vulnerability that SEQ is taking advantage of. Of course I was under the impression that new spawn info was in the clear and was used as the baseline to find the zone in information (but I'm not a developer, so could be wrong).

high_jeeves
03-10-2002, 11:52 PM
Cryonic pretty much has it, except that the new spawn information is encrypted, not plain. (The three encrypted packets are new spawn, zone spawn, and player). They key is currently different per player per zone-in (changes every time you zone). During the WinSEQ days, the encryption would be changed every 3-8 or so days, which would make ShowEQ defective 70% or so of the time. The patch download that was required was small, maybe 3-5 minutes of dialup.. and extremely minimal over broadband of any kind.

--Jeeves

Gungadin
03-11-2002, 03:52 PM
I'm a little confused. On the one hand Cryonic says 'it would take a day or two for the developers to find the new key', but he also says 'everytime you zone the client and server negotiate or exchange a new key'. I believe he maybe means the developers find the new encryprion method?(shrug, i'm guessing)

I see that a 3-5 minute download might be tolerable for most people, so regular changes of that nature might be an option.

Is it possible to automate what the developers of libeq do when the encryption changes (if they can do what they do in 2 or 3 days rather than years that we are told encryption normally takes to break mebbe it could be automated). Thus you get a 3-5 min download then seq had a x min decryption process then back to normal.

Increasing the encryption strength would surely increase the cpu usage on a per packet basis, causing a significant performance hit, so the more regular changing of encryption would be more viable.

Problems may arise though when people are logged on when the encryption process changes. If encryption changes daily instead of at patch times there is a very high chance of this happening, how could they deal with this? this would lead to nasty ingame problems and lots of petitioning.

S_B_R
03-11-2002, 04:29 PM
Well, it's not so much they could rotate the encryption as they could use a completely different method to encrypt the data. In that case the entire libEQ would have to be rewritten, and that could only happen after the devs figure out how the new encryption is working.

There are many ways verant could change the encryption that would thwart SEQ for a good amount of time. SEQ will always be able to break it (because we have access to the client) but it's all about the time required to crack that encryption, or the time required to decypher the encryption scheme.

Also Verant could use a VERY strong encryption to dynamically send the new encryption rotation to the client. Basically using VERY Strong encryption to protect the weaker realtime encryption Key/codec. So they wouldn't need to have a patch to change the encryption, the new realtime codec algorithm could be sent to the client everytime you zone, that data being encrypted by the much stronger and slower codec... It would just make the zone time a little longer.

I dunno if that made any sense or not but it did to me :D

Cryonic
03-11-2002, 05:05 PM
simple:

Oldest Method: No encryption

Old Method: Fixed key for all zones, so it would take the SEQ developers a day or two to get the key out of the client, put it into libEQ.a and distribute it. Would take a day or two longer for WinSEQ to take the libEQ.a and import it into the WinSEQ code and rebuild. The encryption scheme didn't change, just the key.

New Method: Key is generated everytime you zone, so instead of changing a key in the client, they would change the scheme in the client. Will take the developers even more time to find the new scheme and its weaknesses to break it.

Gullork
03-11-2002, 05:12 PM
Interesting. Extremely interesting actually.

I love learning about stuff like this.

But uhm.... lets not give them any ideas :)

Ratt
03-11-2002, 11:25 PM
They already have all the ideas :)

The bottom line is, it's a zero sum game for them. The absolute worst case apocaliptic scenario for SEQ is that we have to have a key snoop on the client machine sending the key each zone to the SEQ computer.

This is the least desireable mass-market scenario, but none the less, as a worst case, SEQ will still work in that manner.

With that in mind, it's really pointless to devote resources to changing the encryption regularly, because in the end, SEQ will still work, and the client will be slower/zoning will be slower and more prone to problems. There's not really much gain and a whole heck of a lot of loss.

Now ... if SEQ gains popularity (like WinSEQ), at some point, it becomes worthwhile to nuke SEQ on a regular basis to hit all the users. As it stand right now, the gain in disabling the SEQ users is far below a financially viable threshold.

Besides, SEQ doesn't really impact the game much anymore compared to the days of yore. It's a glorified map and mana number snoop. Everything else is basically already available in the client, with the expection of MOB locations, which is helpful, but I'd say it's hardly game impacting anymore... who really cares about a level 40 rare spawn in Karanas anymore? Not many people. All the people likely to use SEQ are going to tend towards the Power player profile, meaning they are on level 60 raids, where rare spawn finding is meaningless.

It's useful to a monk for pulling. It's useful for finding corpses ... neither of these are really a major game impact. Even without SEQ, it'll still get done with minimal additional effort.

Anyway ... enough rambling. As long as the SEQ userbase stays below the financial threshold where it's cost effective to disable, SEQ is fine. WinSEQ will shoot it up over this threshold by orders of magnitude, especially when coupled with things like EQW.

high_jeeves
03-12-2002, 12:15 PM
I agree with Ratt completely, the bottom line with Verant (as with any company) is the money. Since ShowEQ doesnt cost verant any money (no users leave because of cheating like it was with UO and AO when it first came out), there is no good reason for them to waste time and money nuking it. This will remain true unless cheating players begins to cost Verant money (which will not happen without a WinShowEQ/EQW type combo where ANYONE at all can run it).

--Jeeves

fgay trader
03-12-2002, 12:39 PM
Originally posted by high_jeeves
This will remain true unless cheating players begins to cost Verant money (which will not happen without a WinShowEQ/EQW type combo where ANYONE at all can run it).


Yes! WinSEQ is the worst idea ever in this respect.

If anything, you guys should make getting SEQ up and running MORE difficult! Let's remove all those helpful How-To's from the board and leave Linux newbs sol :P