To: Paul Fiondella who wrote (30490 ) 2/24/2000 6:39:00 PM From: Scott C. Lemon Respond to of 42771
Hello Paul, > How digitalme opened the door to eDirectory So the first thing that I'll say is that the "pre-Brainshare 1999" digitalme team deserves credit for opening this door. That appears to be the momentum that got things going ... the engineers who worked on last years Brainshare demonstration should be proud ... > So in the process of demoing digitalme, CNN said hey here > is the app we are working on, what's the engine under the > digitalme hood, our database solution is too slow, can > you do something. Yep ... it appears to be that way ... > Along came eDirectory --- Yep ... > user authentication, Whoa!!!!!!! Who said anything about user authentication???? I've read through the articles a couple of times and have realized that there appears to be *NO* user authentication ... just indexed attribute access ... > and profile information stored in the directory at CNN so > that customized info will pop up on the initial screen. Yes ... via object/attribute access ... without any form of authentication from what it appears. > My only question is what did they do about the user login > procedure? There appears to be *no* login procedure ... > Do they still store the user id and password in the > cookie to allow the user not to have to enter both > everytime the visit the site? I don't think so. "Note that CNN uses a cookie as a unique identifier so that each visitor does NOT have to "login" each visit to the site. The cookie is only a GUID (global unique identifier) that is "looked up" in a database and allows the web servers to personalize content based on the returned preferences." My guess is that they are simply saving the GUID as a directory object reference, and then pulling up the attributes of that object using this name. So as of today, I would guess that I could copy your cookie, and see your page ... ;-) Also, notice that they are probably operating completely out of cache memory: "CNN is running Compaq 6400s with 2GB RAM - 1.5GB dedicated to NDS cache (set dstrace = !mb1500000000) and 500 MB for NetWare 5." > Or have they replaced the whole thing with an NDS > solution that has something else at the front end? If you read the technical implementation doc ( novell.com ) you can see that they are using the indexes and access of NDS to grab the information. It seems to say that the "front-line" web servers are all making requests to the NDS servers ... simple indexed look-ups. (The performance is still very impressive ...) Scott C. Lemon