> Greg> Lastly, I would simply like to point out that monoculture > Greg> installations are very dangerous. It's a concept from > Greg> agribusiness.. if you have all one crop, and a virus comes > Greg> along that can kill that crop, your out of business. > >Very true, and this is a terrifically important message to get out. >Not to be pedantic but actually it is a concept from ecology: the >"business", as Greg puts it, can be any system. Diversity makes for >resilience, and vice versa. Okay aleph, it's not a bug but it is a >way we should be thinking. We can think about it, but what can we do about it? Just as in farming there are reasons why we have the monoculture, and just like they buy more pesticides, we buy virus scanners to fix our solution rather than designing another solution. In fact we have even less ability to move away from it than farming. If a farmer bucks the trend and therebye has a crop when the neighbours have none, he has an advantage. If I don't buy CISCO, maybe there will be some time when my router works but the whole net is down with IOS exploits. What will I give my customers with that router while the whole net is down? I can only expect that at some point my router will be exploited while the whole net is up. If I duplicate the router, does the improvement justify the cost? Customers ring if the net is down for 5 minutes, but how much will they pay for 99.9% uptime instead of 99%? If we moved the net to diverse cultures, how should that be configured? In farming the physical seperation of similar crops stops them infecting eachother, but in the net that is only sometimes the case. Taking the Windows Trojan as an example, physical separation stops it propergating by SMD, but not SMTP. Any mathematitions out there like to work on this? it must be an extension to the 4 colour problem in mapmaking. You want no connections between like equipment, but always have network connectivity if one type of equipment is down. It does not look like a type of network I will be building this week. Yours Ian
This archive was generated by hypermail 2b30 : Fri Apr 13 2001 - 14:49:51 PDT