3 ms·

Well, I certainly hope not to. But I'm a paranoid person by nature, and I run a service for users even more paranoid than I am, so I'd like to get some documen

Well, I certainly hope not to. But I'm a paranoid person by nature, and I run a service for users even more paranoid than I am, so I'd like to get some documentary evidence about what happened. I don't trust the DigitalOne people at all and so far everything we know about the raid has come from them. If I have to I'm willing to throw some money into the legal system to get some answers for our users.

patio11 15y ago
You are now running a denial of service attack on yourself. Pretend I have Oracular knowledge about this case. What could I possibly tell you that would be more important to your users than your next three weeks of regular operations? Is "The FBI had a warrant for someone other than you." that information? Is "The FBI did not have a warrant but thought they had exigency, targeting someone other than you?" that information?
- rdl 15y ago
  Yeah, the takeaway from this is "pick vendors better in the future" (I know a lot of great hosting/colo options, would love to have Pinboard...), and "have redundancy ACROSS providers." Don't focus on recovering losses, focus on the future. Buying colo (or especially higher level stuff like managed hosting or even worse, EC2 or PaaS) from a single vendor with multiple sites doesn't protect you from: 1) Them going out of business 2) Them deciding to bounce you as a customer 3) Them getting raided by FBI/Al Qaeda/etc. at multiple sites You ALSO want to make sure the providers you pick don't share much physical infrastructure (e.g. don't use SoftLayer AND Internap at the same physical buildings...). The hassle is it means 2+ sets of support numbers, 2+ sets of provisioning procedures, 2+ sets of billing and contractual relationships, etc., but it's worth it. This is why having standardized provisioning procedures across vendors is awesome (everyone uses x86 now, but having standardized management tools and APIs would be nice, especially for cloud stuff).

patio11 15y ago

You are now running a denial of service attack on yourself. Pretend I have Oracular knowledge about this case. What could I possibly tell you that would be more important to your users than your next three weeks of regular operations? Is "The FBI had a warrant for someone other than you." that information? Is "The FBI did not have a warrant but thought they had exigency, targeting someone other than you?" that information?

rdl 15y ago
Yeah, the takeaway from this is "pick vendors better in the future" (I know a lot of great hosting/colo options, would love to have Pinboard...), and "have redundancy ACROSS providers." Don't focus on recovering losses, focus on the future. Buying colo (or especially higher level stuff like managed hosting or even worse, EC2 or PaaS) from a single vendor with multiple sites doesn't protect you from: 1) Them going out of business 2) Them deciding to bounce you as a customer 3) Them getting raided by FBI/Al Qaeda/etc. at multiple sites You ALSO want to make sure the providers you pick don't share much physical infrastructure (e.g. don't use SoftLayer AND Internap at the same physical buildings...). The hassle is it means 2+ sets of support numbers, 2+ sets of provisioning procedures, 2+ sets of billing and contractual relationships, etc., but it's worth it. This is why having standardized provisioning procedures across vendors is awesome (everyone uses x86 now, but having standardized management tools and APIs would be nice, especially for cloud stuff).