[AdrianB1]

Azure AD may be a decent solution if you have people doing Office and other work all day long, not when you have manufacturing plants across the world in places with poor Internet connections, SCADA and continuity plans that don't involve reliable Internet. If you need that, the complexity of having on premise AD and Azure AD is just cost with no benefits.

[jedieaston]

Wouldn't AD be just as painful in that instance, since replication could be interrupted between other sites by connectivity outages and the local domain controller could become out of sync?

[hug]

Yes and no.

An AD DC can be offline for as long as 60 days (by default, since the actual length of time is covered by AD's tombstone lifetime, which is configurable) and recover just fine, assuming you're not to worried about the intrinsic fact that changes at or affecting that site aren't replicated immediately.

And assuming you plug the site back into the network somehow, under your 60 day limit, AD will largely just keep truckin'. If you've got sites that are offline for more than 60 days at a time due to unforseen circumstances, well, maybe those sites need some other solution.

Spotty connectivity is just fine for AD, though, especially if you're designing things properly.

[2J0]

We pushed federation caches over the signalling channel of ISDN2 as late as 2007. ISDN is the most commonly misunderstood and under appreciated communications standard. During the early nineties ISSN acquired a horrible reputation for incompatibility, but the very nature and value proposition of ISDN remains the most adaptable and capable L0/1/2 communications standard that was intentionally designed to adapt to and enable legacy protocols and future transition requirements. This is why the D channel is capable of transfer of user defined packet payloads up to 16kbps. Commonly dismissed as the signalling channel, why was it designated as "D channel " and the commonly presumed data channels, "B channels"?

The myriad problems and factors that assailed practical IDDN adoption in the early days when a working ISDN interface was a genuine wonder not of miracles in configuration but the capability delivered (we ran applications in advertising that had financial trading style requirements for transactions and the fact that advertising copy was delivered via ISDN was the kind of integration that was envisioned originally)

Everyone is confusing me with their apparent over reaction to the connection of cloud to remote sites. This was what CICS was written for fifty years ago. I feel like I'm discussing state secrets every time I mention CICS. Can anyone recall the killer feature of NT3.1? MTS, if you must have it. MTS is why SQLserver was so easily portable to Linux. I suspect about everyone who reads HN might have a ball with some of the desperately not trendy things we are involved in. Like my uncle dusting his suit vests, smiling he only had to wait thirty years for them to return to fashion. This is a universal interval someone who can explain it will deserve a Nobel I'm sure...

[temac]

ISDN is an edge point to point protocol (although used to get on a network). I don't really see how it has to do specifically with anything related to AD vs Azure (a data link is a data link, you could has well have used an analog modem, or a custom solution involving an avian carrier), neither do I see how ISDN should be considered as anything else than obsolete technology in all regards nowadays, no matter the nostalgia factor.

BTW B channel is for "Bearer", and D is for "Delta" or maybe "Data". (Wild hypothesis: maybe it was initially for "Data" at a time when B was seldom used for things other than digitized voice possibly with a bit stolen for in-band signaling)

On my side I consider that ATM was pretty sweet at a time when IP&co was complete garbage (and still somewhat is), but I know who won.

As for CICS, well, mainframe are still not dead. I suspect they won't, because why would they? But the model is not the same as the modern "cloud" stuff, at all levels.

[LargoLasskhyfv]

"It's dead, Jim." I had that stuff with 128Kbit/s|16KB/s at home up to around 2004, and shudder at the thought of having an upgrade or patch release of any application, or OS over that. Apart from that, every telco everywhere is discontinuing service for that, or has already done so long ago. The whole "ecosystem" of line technology is gone, not produced anymore. Only surviving niche are internal installations where some gateway translates whichever VoIP to some internal S0-Bus, and the long obsolete phones.

That was that. What was your point again in this context?