I manage ten DS 6.9 sp2 servers for a school district. These servers feed about 26,000 computers, and we're about to add another 2000 or so more computers with the opening of 3 new campuses this summer. I
was looking into upgrading to DS 7.1 now that it has PXE, but after hearing how much stuff is missing or completely screwed up, I consider myself lucky that they're still supporting 6.9 with sp4 (I'm about to start testing sp4).
Seriously - taking away remote control? Sure, ds remote is buggy, but we use it all day every day. OK, so NS has pcanywhere. I was totally unimpressed by carbon copy, and was even less impressed by the version of pcanywhere we used to run here in the district (only 1-2 vers old) before we got DS remote. What's the point of having a remote deployment solution if you can't easily remotely manage the machines afterwards? Yes, there are other tools out there on the market (personally i love dameware) but if a remote control tool has always been part of a product, why would you remove it? my only complaint with ds remote in 6.9 is that you can't cut/paste in/out of the window.
Conditions - I use these on loads of my jobs. It really goes well with HII - I can have ONE build job to cover literaly 30 different models (and all their variants) with one image, so the techs in the field (and the service desk) don't have to weed through a massive list of build jobs or potentially put the wrong build job on a machine. For example - some of our machines have TV Tuners in them, so I have the build job check to see if the tuner is there in the hardware list and then install the appropriate tuner software, but if no tuner is found, skip that step and move on to the next one. Without conditions, I'm going to have to make 30 different build jobs, which somewhat negates the purpose of having HII. Yes, I know I will still use the same base image, but from the technician's perspective, the huge leap forward we made last year by implementing HII just got taken away from them.
from other bits i'm reading, it seems aclient/dagent reports less, and pxe provisioning works almost independently from the DS itself, as if integration was deliberately broken (pxe server now names teh machines one thing, then when it boots to production it gets named something else in the console? boy that breaks a whole list of stuff...
our DS system really can (and nearly does) stand alone, completely independant from our three NS servers. The NS (6.x) is being used for inventory, app control, patch mgmt, and a few other things such as deploying our new AV product. Really the only connection the DSs have with the NSs is that the NS servers happen to be our image servers, but if the NSs were to go away, I could just move my images somewhere else and nobody would know any different. They both use the same SQL server cluster as well, but again, the DS doesn't care what the NS does, as they each have their own database.
I can see the benefits of integrating the two platforms, but even in an environment the size of ours, it's not necessary, and don't feel they should require it. In an environment our size, it would be a seriously MASSIVE undertaking, and i can see lots of downtime.
Maybe by 7.5 they will have a more feasible migration path and will have replaced the features they took out. I can respect the fact that they want the core to be solid before they start tacking on extras, but if it's not AT LEAST as good and feature-complete as the previous version, then IMHO it should not even be a RC version - it should be labeled as beta.
I can't see us upgrading to ds 7.x any time soon unless something big changes to convince me otherwise. We might upgrade NS to 7, but won't be using the DS part(s) of it.