The directory services protocols, I’ve ever heard of are LDAP and Active Directory (which is based on LDAP). I know LDAP stands for Lightweight Directory Access Protocol. So, was there ever a heavy or full DAP, or something else that this is ‘light’ in comparison to?
According to wikipedia ldap is named that because it uses less bandwidth than it’s predecessor (which isn’t named). A bit of digging around shows only protocols derived from LDAP, and Novell’s directory service (which was killed off by TCP/IP, DNS, and the various network file server protocols).
As an aside: apart from LDAP, the second main way of talking to AD today is using COM. As a network protocol, that would be DCOM, which, I think, is built on RPC (remote procedure call)
I think that x.500, the directory standard, was conceived as the directory required for x.400, the email service, which was heavier than Simple Mail Transport Protocol. So network protocols (such as LDAP) would have followed from the definition of the directory, not preceded it: indeed I think some early implementations used x.400 for replication.
Not heavier so much as from a completely different universe: X.400 was from the OSI protocol suite, which competed with and lost to the Internet protocol suite, which has SMTP and POP (Post Office Protocol) for email.
These days, people try to pretend the Internet folks implemented OSI or something, and try to forget that OSI was a competing system which is now almost completely dead.
That’s how it’s still taught. Every networking class I’ve been in has included the OSI model but presented it as an abstract model or paradigm that modern protocols sort-of follow. Aside from the TCP/IP suite, I only ever remember hearing about ipx/spx over a decade ago, and even then it was as mostly dead protocol. Well, that and some routing protocols (e.g. ospf, eigrp, rip,etc.), but that’s a world all it’s own. I’ve never heard the OSI presented as an actual set of protocols or working network stack.
I suppose it’s like the difference between an operating system (e.g. dos, windows 95), and a network operating system. Nowadays the concept of an OS that didn’t have a network stack built-in is utterly unimaginable, except as some sort of hellish punishment.
My understanding is that the OSI model is a conceptual model they teach people on how to think about the different layers of networking, from low level (physical) to high level (application/ports and such)
It’s specifically not tied to any particular protocols, like x.500 or x.400
There was also an OSI protocol suite which was a network stack designed to provide a protocol for each layer of the OSI model. This is where x.500, x.400 come in. They’re a proposed networking suite/stack for each layer of the OSI model. That’s what you’re talking about.
TCP/IP won out, and the OSI protocol suite is largely irrelevant, but the OSI model is still taught as a conceptual tool to understand a network stack.
In the early to mid 90s there were a variety of competing protocols and network types in use, and you needed to be aware of them if you were in IT. By the end of the 90s, TCP/IP on ethernet had solidly won out. There were some holdouts, but it’s really going on two decades since the others were in widespread use for ordinary networks.
Because I was always taught it as the “seven layer OSI model”, I always think of a Seven Layer Burrito from Taco Bell when it’s discussed. I’m pretty sure the people who worked hard to create the OSI model would prefer to be remembered as something other than the “Seven layer OSI Burrito”, but that’s life.
… and now I see that Oracle (pbuh) claims that it’s five layers, and separates Data Link from Physical.
Do you kinda get the sense that the Internet People weren’t all that impressed with the notion of layers? That’s only because they weren’t:
The IETF is the Internet Engineering Task Force. They are as close to authoritative on The Internet as any single group can be.
So they teach that the IETF is implementing the OSI Vision. I guess too many Big Important Organizations invested too much in OSI to accept that it was a bust, so now they have to rewrite history to pretend that there were never any OSI protocols, that the Arpanet/IETF people were on board with the seven-layer model from day one, and that the Internet is still an implementation of the OSI model. So it goes.
