Central Services: Difference between revisions

828 bytes added ,  26 November 2003
moved and edited any and all that made any sense about threats of centralized storage
No edit summary
(moved and edited any and all that made any sense about threats of centralized storage)
Line 12: Line 12:


It's usually easier to start central and go distributed slowly over time, but this requires a strong decision to do so right from the very beginning, so that major [[hardware requirements]] and [[protocol requirements]] will take the long term need into account.
It's usually easier to start central and go distributed slowly over time, but this requires a strong decision to do so right from the very beginning, so that major [[hardware requirements]] and [[protocol requirements]] will take the long term need into account.
----
For [[privacy]] and to push forward a [[Consumerium protocol]], however, we want to avoid central store of [[user data]] where ever it is possible, and we want [[label data]] and [[barcode]]s as well to be pushed out as far as the [[Consumerium checker]] as fast as any of it changes.
Centrally stored data is subject to all kinds of diseases and it's pretty much the primary place [[w:cracker]]s will target to get user data.
Since it will be accessed only by [[server software]] probably via [[w:SQL]] it is important that this software be quite careful to authenticate data requests.  Ideally it would log them also for further [[privacy]] protection, and perhaps in [[Transparent Consumerium]] there will be so little privacy lost in publishing these queries that they can all be made safely publicly visible.
9,854

edits