[nycbug-talk] network diversity

[Bob Ippolito]

On Jun 23, 2005, at 11:24 PM, George R. wrote:

> Here's a couple of articles on regarding network diversity.
>
> http://networking.itmanagersjournal.com/networking/ 
> 05/06/22/170205.shtml?tid=32
>
> http://software.itmanagersjournal.com/software/ 
> 05/06/07/180209.shtml?tid=26
>
> There was a nice debate at USENIX ATC in Boston last year about  
> this. . .someone from MS (against diversity) and another guy for  
> diversity.  I only caught part of the debate, but there's a lot of  
> biology metaphors recurring with every argument.  I think there was  
> also something at HOPE last summer on the topic.
>
> It's a solid argument, I think, to say that diversity is better  
> with networks, as specific vulnerabilities only effect certain  
> parts of the network, and are less capable of replication based on  
> uniformity.  But of course, all our love for open standards does  
> potentially open the door to problems in this arena.

Well it's not the open standards that are the problem.  Standards  
don't have security flaws, software does :)

Also, it's interesting to consider that OpenBSD's love of entropy  
(and all of the re-invented wheels) probably makes it quite a bit  
more diverse than other platforms.

-bob