Samba, Patents, and Interoperability

This dude is a True Believer.

The facts are something like this: Samba is a server for the SMB protocol that was part of the IBM PC Network in 1983. If there ever was an IP around it, it expired long ago. There have been third-party servers for SMB since Tandem Computers rolled out the Jeeves server in 1984.

Samba is attractive to Linux and Unix users because the SMB protocol is much more efficient and sensible than the crazy NFS protocol Sun invented in the Age of Dinosaurs that can’t do record locking, but it wouldn’t have become so successful over the past 20+ years if it had been encumbered by IP.

Google has hired another high-profile weirdo, perhaps just to keep Vint Cerf happy in the part of the company with padded cells.

This dude is a True Believer.

The facts are something like this: Samba is a server for the SMB protocol that was part of the IBM PC Network in 1983. If there ever was an IP around it, it expired long ago. There have been third-party servers for SMB since Tandem Computers rolled out the Jeeves server in 1984.

Samba is attractive to Linux and Unix users because the SMB protocol is much more efficient and sensible than the crazy NFS protocol Sun invented in the Age of Dinosaurs that can’t do record locking, but it wouldn’t have become so successful over the past 20+ years if it had been encumbered by IP.

Google has hired another high-profile weirdo, perhaps just to keep Vint Cerf happy in the part of the company with padded cells.

According to Wikipedia, (which as previously discussed could be wrong, but I don’t think it is in this case) Samba…

is a free software re-implementation of SMB/CIFS networking protocol released under the GNU General Public License. As of version 3, Samba not only provides file and print services for various Microsoft Windows clients but can also integrate with a Windows Server domain, either as a Primary Domain Controller (PDC) or as a Domain Member. It can also be part of an Active Directory domain.

In addition:

Samba is an implementation of dozens of services and a dozen protocols, including NetBIOS over TCP/IP (NetBT), SMB (which is also named CIFS), DCE/RPC or more specifically, MSRPC, the Network Neighborhood suite of protocols, a WINS server also known as a NetBIOS Name Server (NBNS), the NT Domain suite of protocols which includes NT Domain Logons, Secure Accounts Manager (SAM) database, Local Security Authority (LSA) service, NT-style printing service (SPOOLSS) and more recently Active Directory Logon which involves a modified version of Kerberos and a modified version of LDAP. All these services and protocols are frequently incorrectly referred to as just NetBIOS and/or SMB.

Are you telling me that none of those protocols could be patent-encumbered? For example, I believe that Active Directory was released in 2000, so it could easily still have patents covering it.

Samba’s goal is to allow a non-Windows machine play nicely in a Windows environment. By necessity, that involves cloning whatever changes Microsoft introduces to its protocols. Unless you’re telling me that Microsoft networking hasn’t changed in 20 years, I would say Samba has something to worry about.

According to Wikipedia, (which as previously discussed could be wrong, but I don’t think it is in this case) Samba…

is a free software re-implementation of SMB/CIFS networking protocol released under the GNU General Public License. As of version 3, Samba not only provides file and print services for various Microsoft Windows clients but can also integrate with a Windows Server domain, either as a Primary Domain Controller (PDC) or as a Domain Member. It can also be part of an Active Directory domain.

In addition:

Samba is an implementation of dozens of services and a dozen protocols, including NetBIOS over TCP/IP (NetBT), SMB (which is also named CIFS), DCE/RPC or more specifically, MSRPC, the Network Neighborhood suite of protocols, a WINS server also known as a NetBIOS Name Server (NBNS), the NT Domain suite of protocols which includes NT Domain Logons, Secure Accounts Manager (SAM) database, Local Security Authority (LSA) service, NT-style printing service (SPOOLSS) and more recently Active Directory Logon which involves a modified version of Kerberos and a modified version of LDAP. All these services and protocols are frequently incorrectly referred to as just NetBIOS and/or SMB.

Are you telling me that none of those protocols could be patent-encumbered? For example, I believe that Active Directory was released in 2000, so it could easily still have patents covering it.

Samba’s goal is to allow a non-Windows machine play nicely in a Windows environment. By necessity, that involves cloning whatever changes Microsoft introduces to its protocols. Unless you’re telling me that Microsoft networking hasn’t changed in 20 years, I would say Samba has something to worry about.

There was a kefluffle over patents and Samba back in 2002:

http://news.com.com/Microsofts+file-share+rule+makes+waves/2100-1001_3-904017.html

Just because SMB is old doesn’t mean Microsoft (and others) can’t play games with IP. I believe that FAT is even older than SMB, yet here we go:

http://news.com.com/Microsofts+file+system+patent+upheld/2100-1012_3-6025447.html

There was a kefluffle over patents and Samba back in 2002:

http://news.com.com/Microsofts+file-share+rule+…

Just because SMB is old doesn’t mean Microsoft (and others) can’t play games with IP. I believe that FAT is even older than SMB, yet here we go:

http://news.com.com/Microsofts+file+system+pate…

Well yes Tim, while the basic SMB protocol is very old all that surrounding stuff is much more current. Taking it all into account, I supposed MS is in a position to get nasty about some of the stuff Samba does. Thus the issue is one for the anti-trust regulators to take on.

One of the interesting uses of Samba is as a replacement for NFS in the *nix world. While that’s not the “stated purpose” of Samba, it’s certainly common practice. There are some policy implications of that, of course.

Well yes Tim, while the basic SMB protocol is very old all that surrounding stuff is much more current. Taking it all into account, I supposed MS is in a position to get nasty about some of the stuff Samba does. Thus the issue is one for the anti-trust regulators to take on.

One of the interesting uses of Samba is as a replacement for NFS in the *nix world. While that’s not the “stated purpose” of Samba, it’s certainly common practice. There are some policy implications of that, of course.

Interestingly enough, I use Samba in our office to connect my Mac to the Linux machine that serves as a file server. One of the great things that Samba has accomplished is to transform Microsoft’s protocol suite into de facto open standards that anyone can implement. Personally, I think that’s a good thing, and I think it would be unfortunate if Microsoft were able to use the patent system to restrict who may implement Microsoft-created protocols.

Interestingly enough, I use Samba in our office to connect my Mac to the Linux machine that serves as a file server. One of the great things that Samba has accomplished is to transform Microsoft’s protocol suite into de facto open standards that anyone can implement. Personally, I think that’s a good thing, and I think it would be unfortunate if Microsoft were able to use the patent system to restrict who may implement Microsoft-created protocols.

Why would that be a shame? Microsoft invented SMB and the related RPC protocols, and they should be entitled to decide who can copy them under the terms of their choice. SMB emerged victorious after battling it out with several alternatives in the marketplace, such as the NFS protocol which is approved by the IETF and native to Unix. This is a case where Open Source went head-to-head with a proprietary system and lost.

Is there something wrong with “to the victor go the spoils?” It’s not easy to design file server protocols (or any network protocols for that matter,) and it’s certainly legitimate to grant patent rights to their inventors.

In the event that MS were to invoke its patent rights and kill open Samba, people like you would simply switch to NFS and enjoy the “benefits” of Open Source design. That’s fair.

Why would that be a shame? Microsoft invented SMB and the related RPC protocols, and they should be entitled to decide who can copy them under the terms of their choice. SMB emerged victorious after battling it out with several alternatives in the marketplace, such as the NFS protocol which is approved by the IETF and native to Unix. This is a case where Open Source went head-to-head with a proprietary system and lost.

Is there something wrong with “to the victor go the spoils?” It’s not easy to design file server protocols (or any network protocols for that matter,) and it’s certainly legitimate to grant patent rights to their inventors.

In the event that MS were to invoke its patent rights and kill open Samba, people like you would simply switch to NFS and enjoy the “benefits” of Open Source design. That’s fair.

Because it’s extremely inefficient for the state to mandate incompatibility among competing platforms. I made the argument in some detail here.

Because it’s extremely inefficient for the state to mandate incompatibility among competing platforms. I made the argument in some detail here.

Tim, do thoughts like this really shape your policy thinking? Come on Tim, there are a lot of “what ifs” in life. This statement says several things you may not be aware of.

it would be unfortunate if Microsoft were able to use the patent system to restrict who may implement Microsoft-created protocols…

1)Tim implies that nobody should be restricted from implementing any MSFT protocols at all. His disagreement with the patent system leads him to act as if its not there. 2)Tim looks at the patent system in isolation, and separate from other business, technical and policy influences. Just b/c MSFT has patents on some technologies does not mean it will only pursue a shutdown of infringing work, in fact MSFT’s recent patent policies have been to share rather than partake in trollish behaviour. 3)Tim has some weird notion of “efficiency” that follows a set of assumptions he often makes w/ tech policy. Tim assumes that because efficiency is a good thing, the more effiency the better, but Tim does not balance efficiency with other important things such as the end product. Also, Tim’s notion of efficiency is that of an end-all goal, and at times seems to be arguing that competition itself is not efficient. 4) Of all the qualities Tim attributes to FOSS and all the criticisms he lends to firms like MSFT, why can’t he explain exactly and in detail how MSFT is hampering FOSS with its patents, and neither does he address the fact that its FOSS’ weaknesses and non-economic viability that limit it.

Tim, do thoughts like this really shape your policy thinking? Come on Tim, there are a lot of “what ifs” in life. This statement says several things you may not be aware of.

it would be unfortunate if Microsoft were able to use the patent system to restrict who may implement Microsoft-created protocols…

1)Tim implies that nobody should be restricted from implementing any MSFT protocols at all. His disagreement with the patent system leads him to act as if its not there.
2)Tim looks at the patent system in isolation, and separate from other business, technical and policy influences. Just b/c MSFT has patents on some technologies does not mean it will only pursue a shutdown of infringing work, in fact MSFT’s recent patent policies have been to share rather than partake in trollish behaviour.
3)Tim has some weird notion of “efficiency” that follows a set of assumptions he often makes w/ tech policy. Tim assumes that because efficiency is a good thing, the more effiency the better, but Tim does not balance efficiency with other important things such as the end product. Also, Tim’s notion of efficiency is that of an end-all goal, and at times seems to be arguing that competition itself is not efficient.
4) Of all the qualities Tim attributes to FOSS and all the criticisms he lends to firms like MSFT, why can’t he explain exactly and in detail how MSFT is hampering FOSS with its patents, and neither does he address the fact that its FOSS’ weaknesses and non-economic viability that limit it.

Tim says: “…it’s extremely inefficient for the state to mandate incompatibility among competing platforms.”

What?

Tim says: “…it’s extremely inefficient for the state to mandate incompatibility among competing platforms.”

What?

What I mean is that rules that give a company the legal right to decide who may make products compatible with its own often leads to a marketplace with mutually incompatible products, with the incompatibility backed up by the force of law. This is inefficient because high-tech products tend to have significant network effects. I explain all of this in more detail in the link above.

What I mean is that rules that give a company the legal right to decide who may make products compatible with its own often leads to a marketplace with mutually incompatible products, with the incompatibility backed up by the force of law. This is inefficient because high-tech products tend to have significant network effects. I explain all of this in more detail in the link above.

Tim, that link above is a bit broader than the current topic. Can you give a concise summary.

Also, Richard has a point that your argument about inefficiency doesn’t really say much. Remember, innovation happens in markets, not vacuums. It tells us nothing if a market space is not perfectly efficient.

Again, you see interoperablity as an end-all goal. It still says nothing that some products may not interoperate with MSFT technologies.

Yes, technology has high network effects. Firms compete for markets, not within markets. Your ideals of perfect efficiency and perfect competition conflict with the very nature of the technology industries.

Tim, that link above is a bit broader than the current topic. Can you give a concise summary.

Also, Richard has a point that your argument about inefficiency doesn’t really say much. Remember, innovation happens in markets, not vacuums. It tells us nothing if a market space is not perfectly efficient.

Again, you see interoperablity as an end-all goal. It still says nothing that some products may not interoperate with MSFT technologies.

Yes, technology has high network effects. Firms compete for markets, not within markets. Your ideals of perfect efficiency and perfect competition conflict with the very nature of the technology industries.

Basically, legal barriers to interoperability between platforms (which is what a patent on a networking protocol would amount to) are wealth-destroying for the same reason that legal barriers to trade between nations are wealth-destroying. My ability to access a Windows server with a Mac, and a Linux server with a Windows box, increases the value of all the devices involved. Legal restrictions on such compatibility destroys that value.

That’s obviously an extremely cursory summary of the argument. I encourage you to read the whole post for the details.

Basically, legal barriers to interoperability between platforms (which is what a patent on a networking protocol would amount to) are wealth-destroying for the same reason that legal barriers to trade between nations are wealth-destroying. My ability to access a Windows server with a Mac, and a Linux server with a Windows box, increases the value of all the devices involved. Legal restrictions on such compatibility destroys that value.

That’s obviously an extremely cursory summary of the argument. I encourage you to read the whole post for the details.

Tim, look up an interview with Bill Hilf from MSFT, where he explains that interoperability between MSFT and FOSS technologies can actually help MSFT’s bottom line. Hence, MSFT has been working with some FOSS entities, and to my knowledge, has not gone after them with its patents.

Sure, MSFT may use its patents to try and deter interop but I doubt MSFT can entirely prevent others from making interoperable products either though legal or technological means. In other words, your patent argument in this post is a big what if.

I don’t think we disagree that much on interop, but damn, stop being so pessimistic!!!

Tim, look up an interview with Bill Hilf from MSFT, where he explains that interoperability between MSFT and FOSS technologies can actually help MSFT’s bottom line. Hence, MSFT has been working with some FOSS entities, and to my knowledge, has not gone after them with its patents.

Sure, MSFT may use its patents to try and deter interop but I doubt MSFT can entirely prevent others from making interoperable products either though legal or technological means. In other words, your patent argument in this post is a big what if.

I don’t think we disagree that much on interop, but damn, stop being so pessimistic!!!

We may very well think that unbridled interoperability is a very good thing, and we may very well think that it’s sound business practice for MS to go overboard to interoperate with FOSS, but does that imply that the law should force them to do so? A lot of that network neutrality debate, such as it was, had to do with non-business-oriented people trying to foist a particular business model off on the network plumbing suppliers. The right to make bad business decisions is inherent in capitalism, just as the right to vote for Al Gore is inherent in democracy. Freedom doesn’t mean much unless you allow others to do things you wouldn’t want to do yourself.

And while interoperability and network effects are nice concepts, they’re also the source of spam, phishing, viruses, and all that sort of thing. Network security is all about limiting interoperability for the personal good. I could have some “network effects” if my WiFi network were unsecured, but the negatives don’t outweigh the positives for me so I don’t run it that way.

MS invented one means to connecting clients to servers over networks – SMB – and the FOSS community has invented another with NFS and its related widgets. At this point it’s obvious that the MS method is superior, so FOSS wants to copy it. I don’t think that’s a natural right, even if it’s arguably good business for MS.

The Samba developers didn’t invent anything, they just reverse-engineered SMB and now they’re acting all entitled to it. Excuse me, but that’s weak.

We may very well think that unbridled interoperability is a very good thing, and we may very well think that it’s sound business practice for MS to go overboard to interoperate with FOSS, but does that imply that the law should force them to do so? A lot of that network neutrality debate, such as it was, had to do with non-business-oriented people trying to foist a particular business model off on the network plumbing suppliers. The right to make bad business decisions is inherent in capitalism, just as the right to vote for Al Gore is inherent in democracy. Freedom doesn’t mean much unless you allow others to do things you wouldn’t want to do yourself.

And while interoperability and network effects are nice concepts, they’re also the source of spam, phishing, viruses, and all that sort of thing. Network security is all about limiting interoperability for the personal good. I could have some “network effects” if my WiFi network were unsecured, but the negatives don’t outweigh the positives for me so I don’t run it that way.

MS invented one means to connecting clients to servers over networks – SMB – and the FOSS community has invented another with NFS and its related widgets. At this point it’s obvious that the MS method is superior, so FOSS wants to copy it. I don’t think that’s a natural right, even if it’s arguably good business for MS.

The Samba developers didn’t invent anything, they just reverse-engineered SMB and now they’re acting all entitled to it. Excuse me, but that’s weak.

We may very well think that it’s sound business practice for MS to go overboard to interoperate with FOSS, but does that imply that the law should force them to do so?

Well no, but I think it’s an abuse of language to say that allowing the Samba team to implement the SMB protocol amounts to “forcing” Microsoft to do anything. The question is whether it makes sense to give Microsoft the legal right to force the Samba team not to use the protocol. Allowing Samba to clone Microsoft’s protocol doesn’t “foist” any particular business model onto Microsoft. There might be good public policy reasons for giving Microsoft a monopoly over its protocol, but failing to give Microsoft such a monopoly isn’t an act of coercion against Microsoft.

We may very well think that it’s sound business practice for MS to go overboard to interoperate with FOSS, but does that imply that the law should force them to do so?

Well no, but I think it’s an abuse of language to say that allowing the Samba team to implement the SMB protocol amounts to “forcing” Microsoft to do anything. The question is whether it makes sense to give Microsoft the legal right to force the Samba team not to use the protocol. Allowing Samba to clone Microsoft’s protocol doesn’t “foist” any particular business model onto Microsoft. There might be good public policy reasons for giving Microsoft a monopoly over its protocol, but failing to give Microsoft such a monopoly isn’t an act of coercion against Microsoft.

I think it’s interesting that you keep side-stepping the question of why the Samba team chose to implement Microsoft’s SMB protocol instead of building a Windows implementation of the FOSS file access protocol, NFS. They certainly could have delivered the critical functions of file server access without stepping into the realm of Microsoft’s protocol at all. What does this fact tell us about FOSS and innovation?

As it stands, Samba is little more than a cheap knock-off of a protocol that Microsoft paid good money (to their contractor, Sytek) to develop. Given that it cost Microsoft money to get the SMB design in the first place, what principle of justice guarantees the Samba team the right to get it for free, “possession is nine parts of the law?”

I think it’s interesting that you keep side-stepping the question of why the Samba team chose to implement Microsoft’s SMB protocol instead of building a Windows implementation of the FOSS file access protocol, NFS. They certainly could have delivered the critical functions of file server access without stepping into the realm of Microsoft’s protocol at all. What does this fact tell us about FOSS and innovation?

As it stands, Samba is little more than a cheap knock-off of a protocol that Microsoft paid good money (to their contractor, Sytek) to develop. Given that it cost Microsoft money to get the SMB design in the first place, what principle of justice guarantees the Samba team the right to get it for free, “possession is nine parts of the law?”

I have no idea whether SMB is superior to NFS. FWIW, the network I helped administer in the 1990s mostly used NFS, with Samba used only to allow the Windows boxes to access the Unix file system. It’s certainly possible that SMB is a superior protocol.

But I don’t see why it matters. “Cheap knock-off” is just a pejorative word for “competitor.” Generally speaking, competition is a good thing. So in the absence of a good argument for creating monopoly rights, I’m going to assume that the best policy is to leave people free to clone each others’ protocols.

I have no idea whether SMB is superior to NFS. FWIW, the network I helped administer in the 1990s mostly used NFS, with Samba used only to allow the Windows boxes to access the Unix file system. It’s certainly possible that SMB is a superior protocol.

But I don’t see why it matters. “Cheap knock-off” is just a pejorative word for “competitor.” Generally speaking, competition is a good thing. So in the absence of a good argument for creating monopoly rights, I’m going to assume that the best policy is to leave people free to clone each others’ protocols.

The technical merits of SMB vs NFS are irrelevant to the Samba team’s aims. SMB is the chosen protocol because it can interoperate with Windows machines right out of the box, with no need to install any additional software whatsoever.

The technical merits of SMB vs NFS are irrelevant to the Samba team’s aims. SMB is the chosen protocol because it can interoperate with Windows machines right out of the box, with no need to install any additional software whatsoever.

“Cheap knock-off” is a pejorative term for an unfair competitor, one who produces a product based on someone else’s R&D or falsely claiming to be something it’s not. Handbags bearing the Gucci logo that aren’t really made by Gucci fall into that category.

If the Samba team said they need to use the Microsoft logo in order to achieve their interoperability goals, you’d be OK with that, wouldn’t you Tim?

Doug, people install software on Windows machines all the time, it’s really not hard at all. Firefox, for example, isn’t an out-of-the-box part of Windows and it’s doing fine.

“Cheap knock-off” is a pejorative term for an unfair competitor, one who produces a product based on someone else’s R&D; or falsely claiming to be something it’s not. Handbags bearing the Gucci logo that aren’t really made by Gucci fall into that category.

If the Samba team said they need to use the Microsoft logo in order to achieve their interoperability goals, you’d be OK with that, wouldn’t you Tim?

Doug, people install software on Windows machines all the time, it’s really not hard at all. Firefox, for example, isn’t an out-of-the-box part of Windows and it’s doing fine.

Installing software on Windows machines is easy enough, but reverse-engineering protocols to achieve out-of-the-box interoperability is even better. Richard’s insinuation that successful reverse engineering is somehow tantamount to trademark infringement is pretty bizarre.

Installing software on Windows machines is easy enough, but reverse-engineering protocols to achieve out-of-the-box interoperability is even better. Richard’s insinuation that successful reverse engineering is somehow tantamount to trademark infringement is pretty bizarre.

Take the example of drugs, Doug. One company invests in research and field trials, and in the end comes up with a useful drug. Another company reverse-engineers the first company’s drug and puts in on the market for a fraction of the price.

According to you, that’s great.

Take the example of drugs, Doug. One company invests in research and field trials, and in the end comes up with a useful drug. Another company reverse-engineers the first company’s drug and puts in on the market for a fraction of the price.

According to you, that’s great.