Apple Releases Rendezvous for Linux, Java, Windows

mblase writes "Apple released yesterday a developers preview of their Rendezvous technology for Windows, Linux, FreeBSD, Solaris and Java. Rendezvous is an open protocol which uses industry standard IP protocols to allow devices to automatically find each other without the need to enter IP addresses or configure DNS servers." Reader xxdarkxxmatterxx adds a link to a story at Macworld about the release."

Comments

For all those that keep asking.....

[CountBrass]

THIS is one of the reasons I'm prepared to pay a premium for Apple kit.

Re:For all those that keep asking.....

[pohl]

Can't speak for the grandparent's logic, but mine is simply that I cast my money towards vendors that encourage openly (and completely, and unambiguously) specified procotols. It effects me because the world is a nicer place when my machine can communicate nicely with the machines of others. It's not that difficult a concept. Some call it "voting with your consumer dollars".

Re:For all those that keep asking.....

[BigBir3d]

And I am thinking the people that worked on it were paid with monies earned from Apple products sold before the release as well.

Re:For all those that keep asking.....

[Dirk Pitt]

Um...

willfully subsidizing these kinds of projects

Do you have trouble dealing in the abstract?

Re:For all those that keep asking.....

[chewmanfoo]

No reason to buy Apple Hardware? Have you not seen this or this?

Surely, you're talking out of your ass here...

Re:For all those that keep asking.....

[Vilim]

Face it, an x86 port of MacOS is never going to happen. Apple is first and foremost a hardware company. Most people don't buy Macs because of thier hardware, MacOS is the selling point, the hardware is just an expensive tax that you need to pay in order to get MacOS. If they ported it to the Intel platform suddenly people could pirate MacOS and run it without paying the hardware tax.

Apples bottom line wouldn't look very attractive

Re:For all those that keep asking.....

[pjt33]

I use an editor called SubEthaEdit because it's the nicest OS X editor I've found. One feature I occasionally use is concurrent editing of a document via Rendezvous. With this release, the SubEthaEdit team might produce a port to Windows soon, which would allow me to do concurrent editing with colleagues who use Windows.

Re:For all those that keep asking.....

[Anonymous Coward]

Well it's actually more than that. It's the control over the hardware and the software that gives Apple the it just works vibe. Apple only has to qualify and QC a small subset of hardware, if Apple opens up to the PC world with the myriad of hardware combinations available, there is no way they could produce the same user experience.

Re:For all those that keep asking.....

[overunderunderdone]

"Insightful" my ass. "Delusional" would be the better adjective. Releasing an OS for Intel would never happen not because Jobs was paid off but because Apple is now, and always has been, a hardware company. A company that makes ~4 billion a year in hardware revenues is not going to abandon that in favor of the few hundred million they make in software.

Apple was "paid off" but the payoff was to drop a lawsuit not to drop development of "rhapsody for Intel" and/or "rhapsody for Windows" which had already been abandoned.

Re:For all those that keep asking.....

[jurv!s]

What he means is that although he's paid a premium for his "kit", he's had the pleasure of using Rendezvous in action for almost two years now... [and trust me, it is a pure delight. It's so pervasive and stable that I almost always forget to mention it as one of OS X's stronger features when proselytizing to the "unwashed heathens" - j/k ]

I just purchased my first new Mac in 3 years [no, my 3year old DP800 is still more than powerful enough to be my main machine, it just wasn't portable *enough*]. I weighed many options and nearly went delerious trying to beat the Apple tax and all their silly pre-configured options. I seriously looked at one of those nice AMD64 laptops for a whole afternoon. But while I would have saved on the kit, I wouldn't be able to run an OS that somehow always manages to stay at or near the top of all the features with little to no hassle to set up and use. After the delerium abated, I settled on a stripped down 15" PBG4 1.33GHz tricked out with 3rd party RAM. I swear that the desire to lick my backlit keyboard in the dark still has not abated...

Re:For all those that keep asking.....

[jeremyp]

Because Rendezvous is currently next to useless in the situation I find myself in where none of the other computers on the network support it.

If it were adopted for Linux and especially Windows I could finally see if it is any good.

Re:For all those that keep asking.....

[Rakarra]

Gotta love that unambiguous, completely open source iPod protocol - that's one of my favorites.

Which protocol are you talking about? The open AAC protocol or the unambiguous firewire/usb drive setup it uses?

Re:For all those that keep asking.....

[foidulus]

Most people don't buy Macs because of thier hardware,
You would be surprised. Outside of the eMac(which is basically just a machine to get OS X), all other apple hardware has some definite selling points.
A portion of /.izens use iBooks/Powerbooks with Linux on them. Why? Because Apple's laptops are a nice piece of hardware. They have decent, if not overwhelming performance stats, they automatically go to sleep when you close the lid, they look nice, certain models don't weigh a whole lot for all the features you get, and they have long battery life.
The iMac has a small form factor, it is very quiet(key among recording studios), and it is sleek, stylish, and dare I say almost sexy. Currently the iMac is a horrible deal, but I would bet on a performance boost and/or price drop beforre back to school season this year.
And I'm not going to start a G5 vs. AMD vs. Intel flamewar, but there are people who believe that the powermacs give the most bang for the buck(but then there are people who disagree, you make your own conclusions)
So Apple hardware(outside of the eMac of course) isn't all that bad. If you are going on pure performance per dollar then maybe it doesn't look great, but keep in mind people have a lot of different needs for their computers.

Re:For all those that keep asking.....

[TheRaven64]

Please will people stop repeating this. Apple is not a hardware company. Apple is not a software company. Apple is a solutions company. Apple products work well because of the tight integration between hardware and software (see the iPod and iTunes for an example). OS X on Intel is not going to happen because it simply would not work. Apple would not be able to provide the same level of integration on x86 (at least not off-the-shelf x86. They could probably release an x86 Mac, although no existing software would run on it without a recompile at the least, and probably some porting for endian related issues in sloppy code).

Look at Solaris on x86. It is very much the poor cousin of the SPARC version, because Sun can control the hardware for the SPARC version and ensure that it is thoroughly tested. Solaris x86 also runs reasonably well on Sun supplied x86 kit, for the same reason. Beyond that, you'd better make sure you pay close attention to the hardware compatibility lists (and, of course, you can't run any SPARC-Solaris binary apps, which are probably the reason you went with Solaris in the first place). OS X on x86 would share another disadvantage with Solaris on x86 - multithreading. Apps on the Mac usually make quite heavy use of multithreading, which is something that x86 does very badly (i.e. context switches on PowerPC are much less costly than x86), so the x86 version would almost certainly run noticeably slower. This would lead people to claim that Apple had released a `crippled' version of the software to drive demand for Macs.

Re:For all those that keep asking.....

[TheRaven64]

Possible, but unlikely. The Rendezvous code is probably a small component of the platform-specific code for SEE. They could possibly port the rest to *NIX using GNUStep, but the Windows port of GNUStep is still in its infancy. If a Rendezvous stack had been the barrier to porting then they could have used an existing open source implementation of ZeroConf (Rendezvous) such as gmdns or howl.

Re:For all those that keep asking.....

[MaestroRC]

That's the difference; on a mac, there is no bios (per se). The settings like that CAN be adjusted with software, software that adjusts the openfirmware settings. And it doesn't require a reboot to happen.

Apple does NOT support running any laptop with the screen closed, because they can very easily (and do) overheat when the screen is close. Especially if you stick that nice toasty laptop into a sleeve in a bag... that's just asking for the thing to overheat. There are utilities out there that let you turn the sleep when the lid closes off, but as I said, if you change that, then the machine is immediately out of warranty from Apple.

Re:For all those that keep asking.....

[AusG4]

He's probably talking about the fact that Apple doesn't offer the specs for the hardware inside the iPod so that he can run Linux on it and listen to Ogg... this is just my guess, but it's probably correct.

Mind, if he had looked at an iPod for a few minutes, he would have noticed that the guts of the iPod run on technology from PortalPlayer

Apple can't open what it doesn't own...

At any rate... you best remember that this is the same type of person who would likely complain that they can't get easy access to the instruction set documentation for the processor in their stereo receiver.... ie .... best just ignore him.

If I had the points left, I'd mod him down as a troll.

Re:For all those that keep asking.....

[SethJohnson]

I did this a few times with a laptop from my previous job. I had it playing MP3s in my backpack as I went to work. It was a sort of mini-jambox. It crashed the hard drive on about the second trip to work... I don't think you want those laptop drives reading while you're running for the bus.. Then again, there's the iPod, so who knows.

Re:For all those that keep asking.....

[oingoboingo]

Apple does NOT support running any laptop with the screen closed, because they can very easily (and do) overheat when the screen is close.

Running a PowerBook with the lid closed is officially supported by Apple, and is mentioned in the instruction manual which is included with the machine. As long as the machine detects an external display and a USB keyboard plugged in, it will run perfectly in lid-closed mode.

I know this is different to closing the lid and then stuffing it into a backpack to use as an overgrown iPod, but the point is that lid-closed operation is not a 'banned' operation because of heat issues. It is very much supported for certain applications.

Re:For all those that keep asking.....

[pigpilot]

I bought an ibook on the basis that OSX would give me a linux like experience without all the hassel you normally get trying to get a wintel laptop running under linux.

Not a great success as OSX isn't what it's cracked up to be. So now I've got an ibook running yellowdog. And as YD knows exactly what hardware it might have to run on it's been more stable than any of my other linux installations so far.

Was the Apple tax worth paying ? . . well the hardware is cute.

Re:For all those that keep asking.....

[nathanh]

A portion of /.izens use iBooks/Powerbooks with Linux on them.

Yup, I'm one of them.

Why? Because Apple's laptops are a nice piece of hardware.

Damn right. Sleek clean lines. Good build construction. Very small. Very light. Long battery life. Everything built-in (Firewire, Bluetooth, 802.11, USB2, Ethernet, DVI), slot loading drive, excellent screen quality. Runs Linux like a charm; almost all the bits are properly supported. CPU is a little gutless but it's more than enough for basics like web browsing, e-mail, office, and quake.

Yes, it was a little more expensive than an equivalent PC notebook. But the PC notebooks are just ugly.

They have decent, if not overwhelming performance stats, they automatically go to sleep when you close the lid,

Yeah, well I don't get sleep on mine, and the video card is nvidia so I don't get 3D either, but apart from that I'm a happy camper.

Re:For all those that keep asking.....

[fmorgan]

I had already said that before, but there was actually a "Mac OS X" for Intel; it was called OpenStep and it flopped.
Mac OS X is actually more an evolution of NeXTStep/OpenStep with a "Mac OS 9 emulator".

Also, does someone remember BeOS? Their business model didn't seem to make much sense.

Some people might buy another OS, but most will just stay with the one offered on the machine.

Lately, I have been thinking that Apple Mac OS X is more like a free "BSD distro" bundled with a computer. Like TiVO is a special Linux distro with a vanilla x86.

Re:For all those that keep asking.....

[gujo-odori]

If it were adopted for Linux and especially Windows I could finally see if it is any good

Rendezvous is pretty good, it works as advertised. You do hit an important nail on the head with your observation though: put a bunch of Win, Mac, and Linux machines together on a LAN and the only really common way of communication they have is SMB. The Macs can talk to each other with Rendezvous, but the Lin and Win boxes aren't joining in.

At work, I run Debian on an almost completely Windows LAN (I may be the only person running Linux on it) and have no problem authenticating to our PDC and browsing the LAN. If you've never tried smb4k, check it out; it's a very nice SMB share browser which mounts shares in ~/smb4k/ for you; if smb4k could be merged with xfe, people transitioning from Windows Linux would hardly know they'd moved, except for the lack of security troubles and crashes :-)

If there were any Macs on our internal network, I'd have to talk to them over SMB, same as with the Windows boxes.

Now, however, the possibility lies open for this to be adopted by Linux, giving a situation where you could throw a bunch Linux and Mac boxes on a network, turn them on, and just let them configure themselves. Or, just a bunch of Linux boxes.

Imagine you're a Linux consultant. You go to a potential customer's office to do a demo with a handful of live CDs that run rendezvous. You pop one into each of four or five PCs, at least one of which has a printer on it.

The machines boot up, they assign themselves IP addresses, find the printer, configure it, find the gateway and configure themselves to use it, and sit their happily waiting for someone to log in.

Then you say to the customer "By the way, the security model is a lot better than Windows, too."

If Linux adopts this and has it implemented well before Longhorn hits the street, it could really help to hurt MS on the business desktop.

Re:For all those that keep asking.....

[aastanna]

Definitly not hardware controlled. I was annoyed when my powerbook running yellow dog linux wouldn't go to sleep when I closed the lid.

Re:For all those that keep asking.....

[bnenning]

Good point, Apple makes most of it's money on hardware. IIRC, that hardware is called the iPod.

Not even close. Apple sold 807k iPods the first quarter of this year, which is maybe $300 million in revenue, compared to total revenue of $1.9 billion. And that was after the release of new iPods and with most of their CPUs being due for an update. Apple can't survive in anything close to its present form as an iPod seller.

Re:For all those that keep asking.....

[bsartist]

One feature I occasionally use is concurrent editing of a document via Rendezvous.

You appear to be mistaken concerning the role that Rendezvous plays in an app like SEE. According to the SEE FAQ, the network protocol used to implement concurrent editing is BEEP.

What Rendezvous is used for is to automatically find other instances of SEE on the local LAN. That's not required for concurrent editing, and in fact SEE allows you to manually specify host names and/or addresses if you need to connect to a machine that Rendezvous can't find automatically.

With this release, the SubEthaEdit team might produce a port to Windows soon

Don't hold your breath. According to that same FAQ, SEE is Cocoa. Unless Apple decides to resurrect Yellow Box, aka OpenStep for Windows, Cocoa apps aren't easily portable to Windows.

BEEP is an open standard though, so you might be able to find a Windows editor that speaks that protocol and works with SEE.

Re:For all those that keep asking.....

[jurv!s]

Mount an iPod and have a look in /Volumes// sometime with Terminal.app . It's all there...

Apple intruding on MS's territory?

[revscat]

Umm... Is it just me, or does this seems to be a little bit of a shot across the bow of Microsoft? Here we have Apple giving something to the community that will add some seriously cool networking capabilities, capabilities the likes of which have traditionally fallen within the realm of the OS itself. At the very least this takes away the ability for MS to use something like this for a "New in Longhorn!" marketing point.

I can't imagine that this makes MS particularly happy, but there's certainly not much they can do about it. Rendevous is seriously a cool technology, and I'm glad Apple decided to release it before MS came up with something similar but incompatible (and, of course, under their control).

Admittedly this argument could be made for Solaris, etc. But I would imagine those communities welcome this addition, whereas I would imagine MS to be a bit colder to the idea.

In any event, kudos to Apple.

Re:Apple intruding on MS's territory?

[harlows_monkeys]

Umm... Is it just me, or does this seems to be a little bit of a shot across the bow of Microsoft?

Microsoft has UPNP, which is supported by pretty much everyone who isn't Apple.

Here's the Linux UPNP stuff.

As far as I can tell, UPNP is the superior technology.

Re:Apple intruding on MS's territory?

[Croaker-bg]

This is a wonderful concept for small to medium sized networks but I can forsee that it will not scale well. If you read into the RFCs a bit it seems that the services location action and the ability to find things without a DNS server rely heavily on the use of DNS via multicast. This is a great idea in principle but the design of a large scale network with correct multicast switching is tough to do from scratch ... let alone reverse engineer your network with multiple flavors of switching gear (cisco, avaya, etc) to handle all the multicast traffic correctly. Sadly, I have to admit that centralized IP based active directory controlled "dynamic DNS" is about the only thing that I have seen that will scale well at the REALLY BIG network level. In addition, I see no mention of the protocol being able to traverse a router. WTF good is a /16 address space if it can't get across a router? Can someone say "DNS helper acl"?

Re:Apple intruding on MS's territory?

[commodoresloat]

I can't imagine that this makes MS particularly happy, but there's certainly not much they can do about it.

Sure they can; they can compete and innovate.

But that's not the way they do things. More likely, they will start by creating their own compatible implementations of Rendezvous until those are incorporated into the OS, and then they'll start subtly breaking compatibility like they tried with kerberos. They'll probably even advertise the new incompatibilities as a great new addon feature to the protocol....

Re:Apple intruding on MS's territory?

[Curien]

Actually, if you'd looked at the RFCs, you might have noticed that Microsoft co-wrote one of them. If you'd actually read it, you might have noticed that they'd been implementing something very similar to it since (at least) Windows 95.

Re:Apple intruding on MS's territory?

[Seanasy]

As far as I can tell, UPNP is the superior technology.

How so?

Is UPnP adopted by any standards body?

Re:Apple intruding on MS's territory?

[fmorgan]

Rendezvous is not something targeted to medium/large enterprises (even if I just heard something about someone at UCLA being able to access all their computers for some "instant-on-grid-calculation with it).
But if you have a home network (TiVo, HiFi with something like AirTunes Express, and iTunes in your PC/Mac), this is great.

No, this isn't a solution for everything, but neither is using a cannon to kill flies.

Re:Apple intruding on MS's territory?

[MrLint]

IIRC HP has had rendezvous support in their printers for a while now.

Re:Apple intruding on MS's territory?

[XMyth]

It's not the concept that is a security hole...it was (is) all the buffer overflow/run exploits in Microsoft's UPnP implementation.

Re:Apple intruding on MS's territory?

[Freedom Bug]

We estimated 6 months of development effort to do uPnP in our embedded network device, and 2 do to zeroconf.

Guess which protocol our device now supports?

uPnP may be technically superior, but more devices will support zeroconf.

Re:Apple intruding on MS's territory?

[falcon5768]

BTW Active desktop was copied from Apple system 6... and there are already a few major and minor companies using Rendezvous most notably HP....

No one is using UPNP that I know of.

Re:Apple intruding on MS's territory?

[mrtrumbe]

As far as I can tell, UPNP is the superior technology.

Many would disagree. O'Reilly wrote up a good piece about the two technologies a while back. Its a good overview of what they offer and the pros and cons of each.

A quick rundown? Zeroconf (Rendezvous) provides a way to discover services and addresses automatically, without address duplication by multiple devices. UPnP does the same thing. The difference is that UPnP also dictates HOW to talk to devices and services, while Zeroconf forces devices to know how to talk to each other.

Taft

Re:Apple intruding on MS's territory?

[outZider]

"Apple's track record in that field as of late hasn't been too great either."

Yeah, because that one security flaw sure was a bitch.

Re:Apple intruding on MS's territory?

[Freedom Bug]

grr, use preview carefully!

That's 6 MONTHS versus 2 WEEKS.

Bryan

Re:Apple intruding on MS's territory?

[Barlo_Mung_42]

"Explain to me what Apple has copied... especially from Microsoft."

Fast User Switching.
Even Jobs admits MS beat them to it.

Re:Apple intruding on MS's territory?

[RetiredMidn]

IIRC HP has had rendezvous support in their printers for a while now.

Yes, and at the risk of ruining a perfectly good karma, I will point out that after trying and failing for half an hour to print a web-based document from my Linux machine on my employer's network printers, I put my PowerBook on the net and started the print job in less than 30 seconds via Rendezvous discovery.

But the really cool thing is that the HP printers on the net show up in Safari's Bookmark bar Rendezvous menu, providing HTML interfaces for printer status and settings.

Re:Apple intruding on MS's territory?

[ad0gg]

Um system 6.0 didn't have anything like active desktop. 6.0 is when multifinder came out. System 7.0 had desktop stickies, but that isn't active desktop.

BTW, most of home routers use UPNP. Netgear, linksys, dlink etc.

Re:Apple intruding on MS's territory?

[csw]

... but neither is Rendezvous. Nor will it be.

Huh. The IETF might disagree...

Re:Apple intruding on MS's territory?

[drinkypoo]

upnp sucks. It has no security, which means that security must be implemented at the device level or neglected - the latter course is linksys' decision, unsurprisingly enough.

I don't know if Rezendevous is any better, but upnp is not all that technologically sexy. It uses http and xml, woop de doo. I'd prefer something a little simpler. Since you can't eliminate the broadcasts, eliminate the http, do everything with UDP, with broadcasts for locating devices. There's nothing wrong with inventing your own simple protocol and it would be cheaper to implement in small devices than xml and http.

Re:Apple intruding on MS's territory?

[pyrrhonist]

Your first sentence is correct ("IETF draft standards" would be more accurate). Your second is not.

You are completely wrong.

Draft Standard is a Standards Track Maturity Level. RFC is a document type.

You were probably thinking of Internet-Drafts, which are not the same thing as RFCs. RFCs must first be published as Internet Drafts. Internet-Drafts are removed after 6 months if they are not approved by the IESG for publication. RFCs are archived.

The zeroconf documents are Internet-Drafts. They are not RFCs.

Re:Apple intruding on MS's territory?

[commodoresloat]

grr, use preview carefully!

Or use it carelessly, and double your karma!

Re:Apple intruding on MS's territory?

[killjoe]

I had fast user switching in linux. Pressing control Fx key allowed me to log in as different users on different terminals. It worked great.

Re:Apple intruding on MS's territory?

[adpowers]

It is really fun when your high school network administrators know nothing about it, so all the Rendezvous enabled printers are sitting on the network advertising their settings menu with no password protection at all. I'm not the malicious type (and I want to cover my ass), so I didn't change anything, but it was a huge security hole.

I discovered this by using my PowerBook at school one day and looking at the bookmarks in Safari. I'm one of the few that brings Apple laptops to school, and probably the only that knows of Rendezvous bookmarks. It is sad Apple develops all this cool technology and most people don't know about much of it.

Andrew

Re:Apple intruding on MS's territory?

[fader]

(Not the original poster, but I'll respond anyway.) It's not like su at all... it's like different terminals. Each virtual terminal is a new login.

What was described with the Fn keys is a bit clunky, IMHO. There's no automatic security on there -- you can switch to another open terminal just by hitting the appropriate button. However, I remember Ximian adding a 'New Login...' option to the GNOME menu sometime around 1.4(?). It performed just like the fast user switching in XP/OSX... you get an XDM/GDM/KDM login window asking for a username and password, and the display on the original login just got locked like XScreensaver does. It was fairly well hidden, so not many people seem to know it was there, but I used it frequently and loved it. And this was before XP came out -- I remember wondering why MS acted like switching users on the fly was such a big deal when I'd been doing it in Linux already.

Re:Apple intruding on MS's territory?

[MrLint]

in my personal experience, good school IT infrastructure is nearly non existent.

So let's see...

[Anonymous Coward]

Rendezvous is an open protocol which uses industry standard IP protocols to allow devices to automatically find each other without the need to enter IP addresses or configure DNS servers.

Following the link to the developer site we find that:

Rendezvous requires that devices implement three essential things. These devices must be able to

allocate IP addresses without a DHCP server

translate between names and IP addresses without a DNS server

locate or advertise services without using a directory server

ok...

Re:So let's see...

[strictnein]

maybe I'm just reading this (and the info) wrong but is Apple just saying: "No DNS/DHCP/Directory Server needed, because everything is a DNS/DHCP/Directory server!"?

Am I confused (most likely) or does that just seem a little silly?

Re:So let's see...

[webguru4god]

Rendevous is some what like making each computer a DNS/DHCP/Directory server, however that's not completely accurate. Rendevous exists as an alternative to those servers for use in environments where either DNS/DHCP is not available, or is not advantageous to configure.

A prime example of Rendevous is two Powerbooks in a cafe, both with Airport wireless. You can set up an Ad Hoc wireless network between these computers, and they will auto configure their IP's and other information so that they can talk to each other. Then open up any Rendevous enabled app and you'll be able to see the other users resources, i.e. bookmarks, printers, music, etc.

So Rendevous is not designed to replace DNS/DHCP, but merely to find a way for network configuration when there is no established network structure. Rendevous also works on networks where DNS and DHCP are available, without any change. This is really the beauty of it, because it can determine what configuration is necessary and do whatever needed to get the computers networked, all transparently!

Re:So let's see...

[dOxxx]

Rendezvous is based on a few underlying services: Dynamic Configuration of IPv4 Link-Local Addresses, Multicast DNS, and DNS-based Service Discovery.

Dynamic Configuration of IPv4 Link-Local Addresses: In the absence of a DHCP server, the machine is able to configure itself with an IP address in a reserved range such that it doesn't clash with other IP addresses configured by other machines on the same network in a similar manner.

Multicast DNS: Multicast DNS Responder services on each machine respond to multicast queries for their DNS information.

DNS-based Service Discovery: Querying for the existence of services (HTTP, FTP, SSH, etc.) via DNS.

So, in a sense, "everything is a DNS/DHCP/Directory server" but only for the information and services provided by that particular machine.

Re:So let's see...

[Twirlip of the Mists]

Rendevous is some what like making each computer a DNS/DHCP/Directory server, however that's not completely accurate.

It's not so much that it's not completely accurate as it is that it's completely wrong.

Rendezvous has nothing to do with DNS, DHCP, or directory services. It's a service discovery framework, and that's all. Here's how it works.

Let's say you've got some program, Foo.app, that has a feature for talking to other instances of Foo.app over the network. Doesn't matter what it is. It could be iTunes music sharing or iChat or distributed compilation or whatever you want.

Without Rendezvous, you'd have to tell your instance of Foo.app where to find other instances. That'd require some kind of setup and some kind of maintenance.

With Rendezvous, your instance of Foo.app sends out a single multicast message when it starts up. That message says that there's an instance of Foo.app available at our IP address. Other instances on the network receive that message and make a note of it. They maintain a list in memory of available services, all automatically, without your intervention.

Does this involve a lot of network traffic? Not really. It requires some, but not much. When an instance of Foo.app starts up, it (1) announces its own presence, and (2) sends out a multicast request for other instances, and the other instances reply. When Foo.app shuts down, it sends out an announcement of its own termination. That's it.

Does this involve terrible security risks? Not really. All Rendezvous does is publish the availability of services that are already running on the network. The responder daemon itself doesn't run with any privileges (on a Mac, it runs as the "nobody" user), and all Rendezvous requests are handled by that one daemon. If something magical happened and somebody was able to get mDNSResponder to run arbitrary code, there would have to be another exploitable security hole somewhere else on the system, because mDNSResponder doesn't have privilege to do anything.

A prime example of Rendevous is two Powerbooks in a cafe, both with Airport wireless. You can set up an Ad Hoc wireless network between these computers, and they will auto configure their IP's and other information so that they can talk to each other.

That's not Rendezvous. That's nothing more than self-assigned IP addresses. When your computer can't find a DHCP server, it self-assigns an address in the 169.254/16 network. Which means any two computers on the same network segment that have self-assigned IPs can talk to each other. This has been around since long before Rendezvous.

So Rendevous is not designed to replace DNS/DHCP, but merely to find a way for network configuration when there is no established network structure.

No, that's overstating it. The sole purpose of Rendezvous is service discovery. That's it. It's independent of network configuration. It works with or without DHCP, DNS, or any other network stuff. As long as you've got an IP address, Rendezvous does its thing.

This is really the beauty of it, because it can determine what configuration is necessary and do whatever needed to get the computers networked, all transparently!

No, no, NO! That's not Rendezvous, that's DHCP. Rendezvous is ONLY for service discovery. Rendezvous doesn't set your IP address or your routing table or your hostname resolution parameters. It doesn't do any of those things. All it does is facilitate service announcement and discovery for your applications.

This is great!

[bennomatic]

Now I don't have to switch to a Mac to have a machine that "just works" on my company's mostly-Mac network!

This sounds like they are getting ready

[great om]

to release more home electronics type products. After all why would they need Rendevous on non-mac platformsm, unless they were planning on selling a networking device that hooks up to the home network? (like, say, a digital video device or some other home theatre component)

Re:This sounds like they are getting ready

[mattjb0010]

to release more home electronics type products. After all why would they need Rendevous on non-mac platformsm, unless they were planning on selling a networking device that hooks up to the home network? (like, say, a digital video device or some other home theatre component)

Like this? Also, I hear portions of XCode are going to be released as open source, anyone got a link? The distributed build stuff is sweet, just tick a box and XCode and Rendezvous do the rest. I heard there's a Linux equivalent to this somewhere, a link would be useful for me right now, thanks.

Re:This sounds like they are getting ready

[Anonymous Coward]

XCode uses distcc for it's distributed compiling, distcc isn't rendezvous aware by default though, so you'll have to configure it yourself.

Re:This sounds like they are getting ready

[chia_monkey]

Wowzers. You could be correct. They just said "hey, check out AirPort Express. You can play iTunes all over the house". And then the dandy new giant displays...not bad for watching TV and such on, eh? And here's a funky little thought...the little widgets that will be released with Tiger. Sure, a nice little calculator here, an iTunes controller there...hey wait...it can control iTunes, maybe a controller for a TV tuner...and your TiVo...and...

Speculation...sure. But we're talking about Steve Jobs and his "digital hub" idea that he announced YEARS ago which finally seems to be coming to fruition. He's getting developers excited about the OS with the widgets, with Rendezvous, etc...lookout MacWorld!

Re:This sounds like they are getting ready

[YouHaveSnail]

After all why would they need Rendevous on non-mac platformsm, unless they were planning on selling a networking device that hooks up to the home network?

- Because in the world of protocols, your "standard" isn't actually a standard unless you can get other people to follow it. Making it easy for others to follow you gives you influence the industry.

- Because Apple would rather live in an open world than in a Microsoft world. (Don't forget, Rendezvous is *not* an Apple invention. It's Apple's name for "zero-conf," and Apple never claimed to have invented it. Apple just made it popular.)

- Because Apple's not selling Rendezvous anyway. They're selling computers, and people will buy Apple computers if they play nicely with others, and if it's easy for others to play nicely with them.

Re:This sounds like they are getting ready

[Elwood P Dowd]

- Because Apple would rather live in an open world than in a Microsoft world. (Don't forget, Rendezvous is *not* an Apple invention. It's Apple's name for "zero-conf," and Apple never claimed to have invented it. Apple just made it popular.)

Right... but the guy who started the IETF working group & made zeroconf happen was an Apple employee at the time, and it wasn't exactly a side project.

Re:This sounds like they are getting ready

[overunderunderdone]

Rendezvous is *not* an Apple invention. It's Apple's name for "zero-conf," and Apple never claimed to have invented it. Apple just made it popular.

Actually, they sort of did. Zeroconf was initially developed by Stuart Cheshire who is an Apple employee with the job title of "Wizard Without Portfolio".

Pseudocode for accomplishing this

[Atario]

for A = 0 to 255
for B = 0 to 255
for C = 0 to 255
for D = 0 to 255
ping A.B.C.D
if (there was a response) then store A.B.C.D in list Q
next
next
next
next
print list Q

Re:Pseudocode for accomplishing this

[JohnFromCanada]

"for A = 0 to 255 . . . ."

How many script kiddies do you think are going to copy that code and try to compile or execute it?

Re:Pseudocode for accomplishing this

[bsd4me]

You forgot to initialize Q to a null list...

Re:Pseudocode for accomplishing this

[bbh]

Sweet! Where do I get a pseudocode compiler! :P

bbh

Re:Pseudocode for accomplishing this

[Anonymous Coward]

Or what, he'll get a pseudocode compiler error? Will his pseudocode have a memory leak or a security hole or something?

Re:Pseudocode for accomplishing this

[Mateito]

This is slashdot.

Why didn't you write it in obfuscated perl? :)

Re:Pseudocode for accomplishing this

[thechao]

You could at least have made it IPv6 compliant...

Re:Pseudocode for accomplishing this

[schon]

Why didn't you write it in obfuscated perl?

Wait, wait, wait...

There's such a thing as unobfuscated perl? Or were you being redundant on purpose? :o)

Everything working together?

[PornMaster]

The only stuff I want to have interconnect is my stuff to my other stuff, not your stuff to my stuff. I'd rather see a simple-to-config protocol than a zero-config autodiscovery protocol.

-PM

Re:Everything working together?

[switcha]

Not sure of the implementations for Linux and Windows, but all of the Rendezvous stuff that I use can be turned off, app by app.

iChat can be set to not look for Rendezvous clients, iTunes can be set to not look for playlists (or broadcast them), Safari can be set to not seek other bookmarks, printing can be set to not seek printers, or the whole protocol can be shut down. I personally only use it for iChat at work, and it never gets in the way or is annoying with anything else.

Re:Everything working together?

[mikrorechner]

The only stuff I want to have interconnect is my stuff to my other stuff, not your stuff to my stuff.

Man, that really doesn't belong here, you know...

I'd rather see a simple-to-config protocol than a zero-config autodiscovery protocol.

Oh wait, you're talking about computer stuff? Well... never mind, then.

Re:Everything working together?

[David McBride]

This is a general purpose service *discovery* protocol. It doesn't define how different devices talk to each other, merely provides a way for them to discover each other's existance.

Knowing another service exists is different from being granted the rights to use that service.

Cool.

[b0r0din]

Sounds like sort of an advanced DHCP. Is this for implementation on an IPV6 or standard IP4 network? Does it include building routing tables between objects as well? Maybe someone can explain it better than I understand it.

Re:Cool.

[MachineShedFred]

If you look at it using a sniffer like ethereal, it's actually multicast DNS, so it will work with either IPv4 or IPv6 as it's a layer-4 thing.

Re:Cool.

[Cajal]

Rendezvous is actually three technologies: link-local IP addressing (which lets machines assign themselves IPv4
addresses without a DHCP server (IPv6 already has this ability)), multicast DNS to find other services on a network, and DNS Service Records to reterieve information about those services.

Apple's implementation of ZeroConf only works on a single subnet. The current draft of the link-local RFC requires that ZeroConf-aware IP stacks reject any IP packet with a link-local address and a hop-count greater than 1.

But, the other 2 parts, multicast DNS and DNS Service Records, can cross subnet boundaries. So, it's possible that you could create a self-updating directory of machines on a multi-layered network, perhaps by using Dynamic DNS along with the other parts of ZeroConf.

An attempt on a Ruby Rendezvous service...

[tcopeland]

...was started way back when.

Since there are now native service implementations available, it'd probably be better to just hook into those using Ruby/DL or some such...

WOW!

[stratjakt]

I cant wait for my network to fill with UDP broadcasts!

This is basically how system-linked xboxes work.

It's cute for little networks that consist of an apple, a printer and an ipod, but it doesn't scale well.

I like my dhcp, that I can control based on MAC addresses.

But kudos to Apple for opening this source. They really had to, you know, one thing they desperately have to overcome is the awkwardness of mixing Mac's and PCs on the same network.

Re:WOW!

[Cajal]

UDP traffic is pretty lightweight. ZeroConf is basically just some ICMP traffic (when the nodes are assigning themselves addresses) and then DNS. Apple's implementation will aggressively cache query results, and the devices incremementally scale back their announcements.

Another nice feature is that nodes can cache the results of other nodes' queries. Since all of the DNS traffic is mulitcast on the local subnet, every node sees every query and every response. Apple's code expolits this to further reduce the need for duplicate queries. It's a pretty nice setup.

Re:WOW!

[yabos]

There's really not a lot of traffic caused by this. There'd certainly be a lot more traffic caused by whatever you will be doing between the computers which you used Rendezvous to discover.

Re:WOW!

[Anonymous Coward]

I cant wait for my network to fill with UDP broadcasts!

This is basically how system-linked xboxes work.

It's cute for little networks that consist of an apple, a printer and an ipod, but it doesn't scale well.

Sorry to be blunt - but how the hell do you know how well it scales? Have you read the relevant drafts? I have - and it's actually amazing how much work Cheshire and Krochmal put into making sure it would be extremely scalable. I don't know hard numbers on what the upper limits on subnet size would be, but I was recently at on a LAN with more than 500 Macs connected with no noticeable effect on the network. Sure there's a limit somewhere, but it's way way way above "a Mac and a printer".

I like my dhcp, that I can control based on MAC addresses.

But kudos to Apple for opening this source. They really had to, you know, one thing they desperately have to overcome is the awkwardness of mixing Mac's and PCs on the same network.

The source has been open ever since they started it. All they're doing now is making easily distributed binaries and SDK's available.

Re:WOW!

[mindstrm]

Multicasts you mean?
Effectively the same in this context.. but still.

A tiny UDP multicast is no worse than the typical windows broadcasts or arp broadcast that happens in a network... and this is just for service discovery generally, not normal net-wise DNS lookups. That means things will be looked up once and cached.
(service discovery could be, by the way "WHO HAS A DNS SERVER"or "WHO IS A NET GATEWAY")

That means, provided a large network has proper multicast aware switches, traffic is minimized.

This is not a replacement for DHCP; it is not meant to scale really huge. it is meant to be able to plug in a machine and have things work in the absence of infrastructure.

Now what I'm wondering is.. what awkwardness is it you are referring to in mixing macs and pcs? We have macs and pcs here, and everything works just fine.

Re:WOW!

[metalligoth]

It's cute for little networks that consist of an apple, a printer and an ipod, but it doesn't scale well

It scales great. I tried the iChat beta when it first came out, and it has Rendevous. I was at Eastern Michigan University, in my girlfriend's dorm. The public beta just came out an hour before, and a girl from across campus popped on my iChat Rendevous window.

My g/f hopped on her iMac. We tried to organize a threesome, but the girl from across campus wasn't up for it. Ah well, at least we all had a good conversation about Apple and their cool technology.

Appletalk for IP

[Anonymous Coward]

Rendezvous/ZeroConf is basically Appletalk for IP. While Appletalk had its shortcomings, it was awesome for setting up small networks. Just plug + play, no DNS/DHCP/etc BS to worry about. Appletalk's gone the way of the dodo, replaced by this (which works on an IP network).

Im ready

[BlindSpy]

I have a feeling this will be implemented into standard Linux use real fast. Having this technology for every platform will really help portability of hardware I think too. This is going to be another one of those things that Windows implements but does it horribly unstable so no one can really use it.

I've found it!

[daringone]

Sweet! My PC just found my Microwave!!!

/home/daringone#setmwave 1m

Microwave set to 1 minute

/home/daringone#startmwave

Your food is cooking.

/home/daringone#

Your food is done.

/home/daringone#

Re:La di da

[ughhgu6]

DHCP is real useful when you don't have a DHCP server on the network....yeah....uh-huh

Apple supports Internet Explorer???

[CdBee]

"Rendezvous technology is now available on Windows 2000 and XP. This preview release includes full link-local support, allowing Windows machines to discover advertised HTTP and FTP servers using Internet Explorer"

Given that Apple today joined the announcement with Mozilla and Opera of open-standards for web plugins it surprises me that their product even suggests the use of Internet Explorer.
I freely admit to hoping, someday, for Safari on Windows and using Firefox until that day (And pls don't reply saying Safari is on Windows in iTunes.. iTMS on Windows doesn't use Webcore, more's the pity.)

Re:Apple supports Internet Explorer???

[jdawg]

Actually, iTunes on Mac doesn't use the WebKit either. The music store is all done with QuickTime, which is why iTunes for Windows also installs QuickTime. For more info, see Hyatt's blog-post about this:

http://weblogs.mozillazine.org/hyatt/archives/20 04 _06.html#005666

DNS-SD

[mabu]

A cursory examination of some of their documents seems to indicate the plan involves what they're calling DNS-SD (DNS-based service discovery) which is a way of encapsulating device id and configuration information within DNS records, and specifically making use of special conventions for TXT data.

If this is the case, it seems a pretty clever and resourceful approach.

Then again, this will make DNS servers the main entry point for discovering information about networks, especially information that might normally not be publicly available.

Personally, I like this approach because far less people have access to manage detailed DNS data and may actually be able to manage these things effectively, but there's also a ton of people out there who have insecure DNS information and adoption of this approach among those admins who haven't secured their networks might create an even bigger security problem.

Re:DNS-SD

[cbiffle]

Close, but with one exception:

The SD component of Rendezvous doesn't use DNS servers per se. Any machine that has services (and that can be any node on the net) advertises these services using multicast DNS.

mDNS is used both to advertise available machines, and the services on those machines. So, essentially, the service-discovery component of Rendezvous is a decentralized, local-link DNS service -- it even uses the DNS protocol (albeit multicasted).

And for the people who keep bitching about UDP broadcasts: first of all, multicast != broadcast. But more to the point, I haven't found Rendezvous to be particularly chatty; Apple claims in its docs to have gone to lengths to prevent it from spamming the network. (The developer docs include some thinly veiled references to NetBIOS.)

itunes, ichat and p2p

[minus_273]

given this is one of the fundamental technologies used for discovery and saring data in itunes, i am surprised it took them so long to release it. It also means that people can write more itunes compatible players (hint to all those linux itunes knockoffs). compatible in the sense that it will appear as a avaiable share and will see other shares on the network. Now all we need is an icaht compatible chat client so that people on a lan do not need a central server to IM :).
It will also be interesting to see how this is applied node discovery in existing p2p systems like gnutella

Re:itunes, ichat and p2p

[cbiffle]

Actually, this doesn't help in creating iTunes-compatible clients.

The Rendezvous protocol specs have been available for some time (it is, after all, a combination of IETF standards). There are compatible implementations for many platforms; on Linux my favorite has been Howl, but JRendezvous is nice too.

iTunes itself uses a proprietary protocol that Apple keeps changing. Parts of it are well-understood (like the music directory portion) but others are not (like the streaming).

Interestingly, Apple is willing to release these docs to developers under NDA. I went through the process but got rejected when they discovered that my product is open-source (even though the iTunes component was to be a binary-only plugin). Apparently they're twitchy about mixed-license developers bending to community pressure and releasing their proprietary info.

Which, with all the GPL bigots I deal with day-to-day, I can kinda understand. Kinda.

Re:La di da

[Dav3K]

How about YOU implement DHCP, if it is so vital to you? Mr. Jobs DID just open the code base, after all, in part for that very reason.

I mean honestly - you whine that it needs to be open so you can code the changes you need, and then you whine when it's finally opened because it doesn't have the features coded for you in advance?

P2P and Rendezvous

[cbelt3]

Darn. Scooped again.. . Note that this technology is NOT Apple's. It's the result of the ZerConf Group. Apple just took the idea and made it 'just work'. And yes, it sounds interesting, and a little scary that all them UDP packets will be whizzing around. OTOH, the reduction of ICANN and other DNS 'gods'is one of the nice benefits of P2P technology like ZeroConf.

Re:P2P and Rendezvous

[overunderunderdone]

Stuart Cheshire, the guy that first proposed Zeroconf and started the ZeroConf group did so as an Apple employee on Apple's dime. I think it's fair to say that it is an Apple technology that they opened up as a standard from the very beginning. This announcement is just that Apple is opening up it's own in-house implementation of an open standard that also started in their labs.

I should point out...

[AusG4]

For anyone who is interested, Rendezvous is Apple's implementation of of ZeroConf

While Apple's Rendezvous overview gives some decent information, the ZeroConf site provides a lot of good technical resources.

Apple really needed ZeroConf as they transitioned to all-IP networking. Although OS X supports AppleTalk, the AppleTalk protocol has clearly seen it's day and the world is clearly moving to IP-only. Previously, when Macintosh machines were largely communicating via AppleTalk, all of the things that ZeroConf addresses were handled by the AppleTalk protocol suite (service discovery, address allocation, etc), and this ease of use that is signature to the Macintosh is important for Apple to maintain.

That said, Apple releasing this code is pretty significant, as aside from this project, there hasn't been much use of ZeroConf in the wild.

Re:I should point out...

[kdogg73]

Get out! Your ZeroConf is run by no other than Stuart Cheshire, the creator of the classic network game of Bolo. Nice.

Re:I should point out...

[MasonMcD]

Also note that Stuart Cheshire is the brainchild of ZeroConf, and he works for Apple, so Rendezvous isn't some hobbled, second-hand implementation of ZeroConf, but from the horse's mouth.

Here's the history of ZeroConf:

The initial seeds of Zeroconf started in a Macintosh network programmers' mailing list called net-thinkers, back in 1997 when I was still a PhD student at Stanford. We were discussing the poor state of ease-of-use for IP networking, particularly the lack of any equivalent to the old AppleTalk Chooser for browsing for services. I proposed that part of the solution might be simply to layer the existing AppleTalk Name Binding Protocol (NBP) over UDP Multicast.

Rendezvous for Pocket PCs

[drazvan]

We've launched this only a few weeks ago: Pocketster. It contains an implementation of Rendezvous for the Pocket PC and it also gives you wireless filesharing capabilities (we have a new version coming out on July 6th). It's free, so give it a try if you want (that is if you have a Pocket PC). Also, you might want to check JmDNS (Java version of Rendezvous) and Howl for a Windows implementation. Razvan

Rendezvous really helps laptop users

[Infonaut]

If you're like me and you find yourself using your laptop at client locations all the time, plugged into their network via Ethernet or a WiFi connection, Rendezvous is great.

Several times I've had the need to print something while in an unfamiliar network. It takes just a few seconds to find and send a job to a printer using Rendezvous. At first it seems ludicrously easy, like it won't actually work. But it does.

In a laptop-centric world, Rendezvous makes life a lot easier.

Netbios

[ad0gg]

Microsoft has netbios as well which announces computers/printers on a network by UDP broadcasts. So why do we need this software again?

Used Rendezvous on Linux before

[weinford]

This is not really new. There was an article in the german IT magazine c't recently, which resembled mostly what is in several hints like this one. It's about using Rendezvous on a Linux box combined with a streaming server for iTunes. Works great and is real good fun!

Re:Appletalk

[Anonymous Coward]

The biggest problem with Appletalk were the stupid cables which used a DIN type connector, which would never stay connected.

Which had nothing to do with Appletalk (a bundle of protocols) but a lot to do with Localtalk (a hardware standard).

UPNP vs zeroconf/rendezvous

[ashpool7]

I wouldn't call UPNP "superior" by any stretch of the imagination.

Comparisons have been done. I'd rather have low traffic and better service separation vs the "use-http-for-everything" strategy.

Re:UPNP vs zeroconf/rendezvous

[Precipitous]

The link is a very interesting article - more interesting than the initial story.

To me, it looks like the largest drawback to UPnP is that it defines too much and becomes inflexible. While the current implementation of Rendezvous is directed at home networks and networks without much infrastructure, I can't see why it couldn't scale out. If / when it scales out, it will intrude on more than just UPnP. It could also kick butt all over MS's UDDI for web services. Here's a scenario for which I could profitably use a more scalable Rendezvous type functionality, where neither UPnP or UDDI would work well:

A service gets a name, independent of the machine. Clients of all types find and connect to the service. For example, we've got a critical Job Status service, that collects information about myriad automated jobs so that the staff responsible for a set of jobs can quickly check if any of their jobs are misbehaving.

Say the computer running the Job Status service blows up and rolls over to a different device (or we painfully restore it on another device). Certain fancy expensive data base servers handle this rollover smoothly - but not my home grown application. I get it almost for free with Rendezvous (expect moving the service). Because the client connects to a service name, it finds the new service seamlessly. No configuration file push, no changing C-Names in the active directory (which requires arguing with about 4 departments in my company) . Just bring up the same service name on a new device.

Rendezvous could apply to any service - not just web services as with UDDI. Also unlike UDDI, there is no need for a single point of failure (the server with the UDDI directory). Unlike UPNP, I don't have to jump through hoops to describe my service, or attempt to conform it to an existing specification - and the current ones are really focussed on devices. I don't really care about describing my service in grand detail. I can assume an application designed to work with knows how it works.

The main thing missing from Rendezvous for this scenario is scalability. Rendezvous could solve this easily by stealing the controller model for UPnP. Put up 5 or 10 controllers on our 10,000 device network. Each client knows about a few of them. You can handle the load and don't have a single point of failure.

Re:UPNP vs zeroconf/rendezvous

[harlows_monkeys]

I wouldn't call UPNP "superior" by any stretch of the imagination.

Comparisons have been done. I'd rather have low traffic and better service separation vs the "use-http-for-everything" strategy

You are focusing on the part that doesn't matter. I suspect that in real life, you'd have a hard time finding any performance difference due to the heavier traffic. It's lost in the noise.

The part that is important is specifying the commands and data to/from devices. We learned this lesson back with SCSI-1. When you leave important things up to each vendor (like the way Rendezvous leaves stuff in unstructured text fields for the vendors to define), you end up having to build into your host software a bunch of vendor-specific knowledge. You end up not being able to just go buy and plug in a printer or scanner or whatever and have it work, because the vendor hasn't released documentation to the Linux developers yet.

Notice how much cleaner everything was with SCSI-2 due to having all the important commands specified in the standard, so that you could have generic disk and tape and other drivers that utilized all the device capabilities.

It seems to me that Rendezvous is making the SCSI-1 mistake, which was understandable for SCSI-1 since they didn't have anything to apply hindsight to. The Rendezvous people should be able to look at SCSI-1, though, and see the importance of complete device specifications.

Re:UPNP vs zeroconf/rendezvous

[Anonymous Coward]

The part that is important is specifying the commands and data to/from devices.

Which already exists as other protocols. It's better to separate the device/application-specific stuff from the transport. We learned that lesson from IP.

We learned this lesson back with SCSI-1.

No, we didn't. Your example ignores the fact that protocols already exist to perform these functions. Why reengineer IPP when it already exists?

A lightweight approach is best, which is what Zeroconf provides.

Re:UPNP vs zeroconf/rendezvous

[janbjurstrom]

Great link, thanks. From that O'Reilly article:

Branded by Apple as Rendezvous (the same way that IEEE 802.11b became Airport and 1394 became Firewire), Zeroconf is designed to bring the "It Just Works" Apple swagger to IP [...] networks.

I wonder - and forgive my ignorance if I'm asking/stating something untrue/stupid: if Rendezvous is "just" an implementation of Zeroconf, why all the "Kudos!" to Apple for "giving back to the community" (not you, ashpool7, but in this discussion)?

Because, I don't see people praising Apple for 802.11x ("Airport"!) or IEEE1394 ("Firewire"!) in quite the same way...

As I understand it, they didn't come up with the underlying technology(?) but rather made an implementation. And now, under their own license, release that implementation - source code to registered users. And again, if I understand this correctly, there are already other Zeroconf implementations, apart from Apple's?

I'm not trying to troll or flame here, but from what I read in the article ashpool7 linked to, I honestly don't understand why Apple are made out to be such heroes. Anyone care to explain?

Re:UPNP vs zeroconf/rendezvous

[tupps]

From what I remember this technology either came from Apple or was given a major helping hand. I beleive that the head of the ZeroConf working group is an Apple employee.

Re:UPNP vs zeroconf/rendezvous

[janbjurstrom]

You're right, thanks. Upon further reading, I came to realize this too (replies to this post and the Zeroconf page).

My bad, Apple does seem to rock in this case :)

Apple does solid software

[quadra23]

Apple continues to surprise me with their interest in designing software that is compatible for the Windows platform in addition to their own (and in this case, also Unix and GNU/Linux). While from Microsoft, they have typically steered to their flagship products and rarely ported them (with the exception of Office and IE) to other OSes.

Even as a frequent Windows user I have great respect for Apple and find their software for Windows actually crashes less then Microsoft made software(!). In addition, they are rarely so deeply entrenched in the OS that if you wanted ot change extension preferences it doesn't fuss as much.

I'd be interested in trying out this new technology and I'm sure it will make it big hit on all network sizes. Good thing for Apple that they released the specs before MS could claim any competing service! Let's all give some positive input to see this software hit new limits!

Confusion.

[FreeLinux]

This "protocol" basically combines several different pre-existing protocols. It seems most like a combination of Auto-IP addressing and SLP. The thing is that OS X and several other OSes including Netware and Linux already use SLP. Now Apple wants everyone to change again?

Here's the really strange part. With SLP, unless you do static configurations, the requests are sent via multi cast. This creates an issue because most routers are not configured to handle multicasting so the SLP scope is limited to the local segment.

Apple's new protocol relies on multicast DNS where, DNS requests are sent via multicast. But, the problem with multicasts being restricted to the local segment still exists. This means that Rendezvous offers no clear advantage over SLP, an already defined and implemented standard. So, why should anyone adopt Rendezvous?

Re:Confusion.

[Chucker23N]

Rendezvous is three things:

- automatic link-local IP addressing for cases where DHCP fails (like APIPA)
- multicast DNS for announcing device names (.local domain)
- service announcements and discovery via DNS-SD

Mac OS X also supported SLP, but Rendezvous / ZeroConf is clearly the more comprehensive technology, as several projects (such as GNOME) are actually moving *over*.

Re:Confusion.

[robertchin]

The current rendezvous stack in OS X doesn't support this, but rendezvous directory lookups are supposed to be able to be advertised by a normal DNS server. That is, a service should be able to be advertised through BIND, so that if a multicast DNS lookup fails, the DNS lookup will work. Although I'm not sure of the precedence -- whether singlecast or multicast DNS takes precendence for a lookup. So rendezvous and traditional DNS are designed to work hand in hand.

Rendesvous, Tiger, and NT

[chia_monkey]

And this from Apple's website:

NT Migration Tool Tiger Server makes it a snap to upgrade your aging Windows NT network to a Mac OS X server. The new NT Migration Tool automatically extracts all of your user and group account information from an existing Windows Primary Domain Controller and moves it into Open Directory. Tiger Server can then take over as your Primary Domain Controller for your Windows clients and even host your Windows users' home directories, group folders, roaming profiles and shared printers.

So they're making it easier for NT users to migrate their network over to Tiger when it is released. And now this Rendesvous news. Sounds like Apple is quite serious about wanting to be a player in the enterprise server market if you ask me.

wrong

[Anonymous Coward]

UPNP is as old as zeroconf and is no more mature than it. Zeroconf is clean and pluggable. It's the published services that may or may not be mature, such as finding web servers, other people to chat with, or another machine with distcc.

People are bashing it here because zeroconf is a small traffic (multicast DNS) linux/OSS technology and UPNP is a high traffic (SOAP aka XML over HTTP) microsoft technology designed to push people to using SOAP everywhere.

zeroconf is rendezvous and vice versa. don't forget

Some links

[Midnight Thunder]

Here are some links that might be of interest:

• jmDNS - Java implementation of mDNS - the technology that makes up Rendezvous. It was formerly known as jRendezvous by Strangeberry
• Zeroconf - more about the technology that Rendezvous implements

Bzzzt!

[MachineShedFred]

"Both of then do support the IPv4 link local address (169.x.x.x) and thats it"

I don't know if you are saying that's all that is the same between them, or if that's the only IP range (which isn't even right - it's 169.254.x.y) that they support.

I have Rendezvous working on three different networks of different scales, all of which are using DHCP to allocate addresses in the 10.254.254.x, 192.168.50.x, and 172.18.x.y ranges respectively. I can go to a Mac and try to connect to afp://server.local on any of those networks, and it gives me an authentication dialog.

MDNS and DNS based service discovery are not bound to any IP range, as they are layer 4 services, and IP is layer 3.

Link-local addressing only exists in Rendezvous for when you need it, not for a requirement.

Re:how does this differ from current projects...

[saddino]

I thought this was already open, how does this differ from apps that have been out for some time like HOWL

Rendezvous, like HOWL, is an implementation of Zeroconf.

This sounds allot like apple's dashboard, and how it's a direct rip of Konfabulator!

FYI, Rendezvous was released to the public July 17, 2002, almost a year before HOWL (June 10, 2003). Apple released it open source, so it's not surpising that they now have released cross-platform implementations.

Mod me redundant

[bahamat]

because I'm probably not the only person who downloaded the rendezvouz code months ago, compiled it and have been running it on their linux box.

Rendezvouz enabled clients on my home network will find my linux box available over rendezvouz for AFP, FTP, SSH, HTTP and IPP.

Mac users will feel (and have felt for quite some time) right at home on my network.

Re:Mod me redundant

[bahamat]

I haven't set up NFS, but it should be just like the AFP stuff. Apple says that AFP, WebDAV, SMB and NFS are all browsable by the Finder.

Anyway, if what Apple is giving out now is the same as what I downloaded a few months ago it's really pretty easy.

I wrote up a quick and dirty Linux Rendezvouz HOWTO-MINI on my website. If you're moderately familiar with UNIX you should be able to use that as a base to set up NFS or anything else you like.

Re:Umm...

[Twirlip of the Mists]

You're not mistaken, but what's new here is the C library for talking to mDNSResponder. That wasn't there in the POSIX example code before, which meant you were SOL if you wanted to actually build Rendezvous support into your application. You could run a Rendezvous proxy on a POSIX system; I set that up at a friend's place. They had a network that was about 75% Mac and 25% SGI. I set up Rendezvous proxies so their IRIX machines would announce their FTP and SSH services via Rendezvous so the Mac users would be able to make use of them more easily. But that was as far as you could go without heavy, heavy lifting.

But now that's no longer a problem.

The only reason for Apple atempting to break DAAP interoperability that I can think of is to enforce the 5-connections maximum present in iTunes.

No, it had to do with stealing. DAAP has to be encrypted to prevent people from being able to easily steal music with iTunes sharing.

They beat microsoft to the punch.

[karmatic]

Microsoft has been in alpha now for a bit with a product called "Windows NCD Technology" (I'm a tester). This is Apple's shot across MS's bow.

From the alpha page, Windows Network Connected Device (NCD) Technology is a comprehensive set of Windows technologies that allow devices on a local network to discover, communicate with, and control each other.

Apple did have an OS for Intel hardware

[Anonymous Writer]

Around the early 90's, Apple developed a prototype operating system called Pink, which they later spun off into a company called Taligent, with the help of IBM and Hewlett Packard. Here is an old article about it and why it was canned. I remember reading something about how former Apple CEO John Scully gave a demonstration to some people of what looked like the Macintosh operating system running on a PC. As I recall it could run on both Mac and PC hardware platforms, and was designed in such a way that programmers could create programs that ran on both platforms through object-oriented programming. I purchased a book on it ages ago in which it was described as an "application system". It was meant to be a true cross-platform operating system.

Re:Apple did have an OS for Intel hardware

[Anonymous Coward]

There were actually 2 different projects. One, "Star Trek" was a straight-forward port of MacOS to x86 hardware.

The Taligent project became this IBM thing called "WorkplaceOS" which was supposed to run both OS/2 and MacOS "personalities" from the same core operating system. Part of it came close to shipping with OS/2 for PPC.

Re:Apple did have an OS for Intel hardware

[PygmySurfer]

Even later than that, early versions of Rhapsody (which became OS X) also ran on both PPC and x86. And of course, OpenStep ran on PPC, x86, AXP, PA-RISC, and who knows what else.

Re:Apple did have an OS for Intel hardware

[Anonymous Writer]

There must be something going on behind the scenes where Apple and Microsoft have made a deal to keep their operating systems on separate platforms. If Apple released an OS for the intel platform, Microsoft could probably easily release a version of Windows for Mac hardware. That would be a head-on confrontation worth watching.

Ironically, when Pink came out, IBM still had leverage with OS/2, so they still had some influence over the intel platform. If Apple and IBM both adopted Pink, they would probably have had an opportunity to curb Microsoft's dominance, compared to the chances of that happening today. Apple probably didn't release Rhapsody for intel because they knew it was too late. Maybe they just used it as a bargaining chip, the way Microsoft uses the threat of discontinuing Office for the Mac.

It's a shame to see corporate politics get in the way of producing better software. Pink, Rhapsody, Java... all the promises of cross-platform programs have never happened. Consumers should be able to go into a store and buy an application that would work on any platform, but as usual, benefits are skewed in favor of the corporatioins rather than the consumer.

Re:Apple did have an OS for Intel hardware

[tbone1]

There must be something going on behind the scenes where Apple and Microsoft have made a deal to keep their operating systems on separate platforms.

No there doesn't. What you see as a conspiracy is actually two companies with different business models.

Apple makes money from selling hardware. This isn't strictly true, they're more of a solutions company than a hardware company, but selling the hardware is where they make their cash. They probably make a bit of cash from their software, but the margin must be small, given the size of their market, and they use their great software to push their hardware. Therefore, it is in Apple's best interest to make their hardware different so no one can run their software on other hardware. That keeps their margins high, and moving to the commodity platform of x86 would mean people would run Apple's low-margin software on low-margin hardware made by others, and Apple's bottom line would take a pounding.

MicroSoft, on the other hand, makes money off software, not hardware. As of 2002 (I don't know about the latest financial report), MicroSoft makes money in three areas: their OS, Office, and the Mac Business Unit. The MacBU is tiny compared to the first two. Therefore, it is in their best interest to write the OS (and per corallary Office) for the commodity hardware, which is x86. Writing for other hardware would increase their overhead, and for what purpose? People who want Windows will get an x86 machine to run it. It's not in MicroSoft's best interest to have to support multiple platforms. Heck, they have a rough time keeping their x86 OS secure and relatively bug-free. So they plug along with x86 until it is no longer the preferred chip.

It's probably more complicated than this, of course; I've heard of agreements and contractual obligations between Apple and IBM/Motorola and MS/Intel that reinforce this seperation, but it is probably more reinforcement than anything.

Re:La di da

[gamgee5273]

Not the same thing... but, then you wouldn't be an AC if you wanted to be taken seriously.

ZeroConf means a visiting professor walks into a lab at a university and can automatically print. There's zero configuration.

It means an iTunes user can broadcast their library on the network and another iTunes user can pick it up with no problem. There's zero configuration.

It means I can open iChat, not go onto AOL's network, and see my coworkers down the hallway with zero configuration.

It means I can share a workgroup document we are editing in SubEthaEdit and easily invite coworkers on the LAN. There's zero configuration

And now it means that non-Mac users can start getting in on a lot of the same stuff.

No parent was right, you are wrong.

[overunderunderdone]

No, you are wrong... or more correctly only partly right. Rendezvous is Apple marketing-speak for zeroconf which involves, to quote the zeroconf website:

1. Allocate addresses without a DHCP server.
2. Translate between names and IP addresses without a DNS server.
3. Find services, like printers, without a directory server.
4. Allocate IP Multicast addresses without a MADCAP server.

You are quite accurately describing point 3 whereas the parent was describing points 1,2 and 4. But ALL FOUR are rendezvous/zeroconf.

Re:Smells like WINS for Mac

[2nd Post!]

Why do you think this will be a noisy, chatty, protocol?

And Rendezvous, if I recall correctly, uses multicast DNS, icmp, and a couple common ones. What it does is piggyback on top of existing protocols.

Which means you would have to block existing protocols...

Use zeroconf to find the router

[ashpool7]

zeroconf enabled DHCP server can point you in the right direction. Zeroconf outside of local networks makes no sense. When you ask for all the local printers, you don't want to get every one on the entire frickin internet.

For enterprise wide networks, you zeroconf/rendezvous acquire a DHCP server and a Directory server. From there, they will point you to the rest of the services in your enterprise *outside* of your local network.

Correct multicast switching is not a problem. Do you personally mess with something that intentionally messes up the broadcast address in TCP/IP?

Re:Smells like WINS for Mac

[gerardrj]

When we were all on shared 10b2, such complaints were completely justified.
When we were all on shared 10bT, such complaints were arguably justified.
When we are all on switched 100bT, such complaints are arguably unjustified.
When we are all on switched 1000bT, such complaints are completely unjustified.

Maybe there is a bit of a problem if your entire network is run on 802.11b and you have dozens of Rendezvous enabled nodes out there.

Rendezvous is not especially "chatty". I think nodes broadcast a small packet at most every 30 seconds to advertise their services or at least availability.

Re:Smells like WINS for Mac

[Chiminea]

We have a couple of hundred Macs and I can see a lot of them coming up and going down just by turning on my Rendezvous enabled iChat. They have to broadcast when they make any changes. Add a couple of hundred happy Windows users into this (running god only knows what kind of Rendezvous enabled apps) and we have even more wasted bandwith. I am an old school "save the bandwith for real work" person. Your comment about wireless was something I hadn't thought about, all our APs are indeed 11b. //sigh

Re:Multithreading

[TheRaven64]

I've not had huge problems with multitasking on a Mac, since thrashing has been a problem whenever the machine is heavily loaded (the only thing I do that really taxes this machine is video editing, and that require a lot of RAM, meaning that when I switch to anything else the amount of paging that has to take place means I really don't notice inefficiencies in the multitasking). On the other hand, I've never had problems with (for example) iTunes skipping when the machine is under heavy load. The problems you are experiencing with the current scheduler likely stem from it not prioritising I/O-bound processes enough. I don't know if the scheduler in OS X comes from the Mach or FreeBSD side, but either way it closely models the 4BSD scheduler. Hopefully 10.4 will import the ULE scheduler from FreeBSD, which is O(1), prioritises I/O bound processes and (unlike the Linux O(1) scheduler) properly respects nice values.

Oh, and I wasn't talking about multiprocessors in particular. A context switch on a single-CPU PPC machine is roughly equivalent to a function call in terms of overhead. On an x86 system it is about an order of magnitude more. This makes multithreading very expensive on an x86 system (one of the reasons why things like games don't often use it).

To all the naysayers.

[mindstrm]

- This is not Appletalk. IT is new.

- This is open

- The *key* feature is the mDNS system.

- Yes, it does automatic IP allocation if there is no DHCP server.. so does Windows (though apple is much faster at it for some reason).

- mDNS is not to be confused with "The global DNS system" that you use to lookup Address records, etc, though it can do that. mDNS is DNS adapted to multicast, for service and host discovery. HOw?

- Instead of querying a DNS server, you query a multicast group (the link-local group in this case) and say "Who has a webserver?" or "who has ssh?" or "hey FOO, what is your IP?" or.. more importantly "Who is a real internet DNS server?" or "Who has an internet gateway?". "Who else is running itunes?".

- a machine joining the network will broadcast once, to send out that it has joined, and what services it has, also via mDNS... so anyone listening can update their caches, etc. The opposite happens when it leaves.

This does not create an extra burden of traffic. Previous to this, most protocols that need to find something in the network do so by rather rude broadcasts.. and usually generate quite a bit of unnecessary traffic.

YES, having a set infrastructure, DNS servers, DHCP, etc, and using DHCP to hardcode everything else, avoiding the need for local network discovery is more efficient. The point is, this works very well WITHOUT any infrastructure.. like 10 guys sitting in a conference room with wireless cards and no servers... or 3 guys on the bus. etc.

NO, rendezvous does not grant ACCESS to your computer.. it merely discovers advertised services... much like an X browser can find a bunch of remote X desktops, or windows TS can find all the terminal servers in the network, or the "network neighborhood" list is populated in windows. It's just a more elegant, scaleable approach.

- mDNS is *not* dns... it is mDNS but if you understand DNS you will understand mDNS. They chose to not make a new protocol, and instead adapt an existing one... which makes it much easier to learn and work with.

fancy things for fancy setups

[ashpool7]

A zeroconf DHCP server could do fancier things such as preventing a rogue DHCP server from taking over (see previous Apple security article).

You seem to be confusing zeroconf as a replacement for something more "sophisticated" such as a directory service. This is not the case. A network admin benefits from rendezvous in the following way:

* A machine, new from the box, goes online
* Rendezvous detects the DHCP server and Open Directory (Apple) server on the subnet
* The machine connects to the directory server and receives its configuration from the directory after the user provides credentials.
* The machine is now configured and can access all resources in the enterprise. No network administrator intervention was required.

Zeroconf merely helps computers find the higher-powered services you're demanding. This is how Apple is leveraging the technology.

If your WAN is not subnetted over those slow links, I would think that you shouldn't have designed the network that way to begin with. Seems like it's just asking for trouble. If you *have* to have them talk to each other, the Directory Server can do that by making the disparaging services visible to both sides and rendezvous can help your machine find it.

it doesn't seem to work?

[rollthelosindice]

We have an internal network here at work that is about 80%/20% mac/pc. The main laser printer is attached to a mac and shared over rendezvous. My Mac discovered the printer instantly and i print to it all the time. My PC sitting right next to me had this developer preview installed yesterday, and I immediately fired up the rendezvous printer setup wizard. Only to discover that it told me there were no rendezvous printers on the network.