Cisco Subnet An independent Cisco community View more

Cisco doesn't contribute nearly enough to open source

Cisco has created some tantalizing open source projects but could, and should, do more.

Cisco's website proudly proclaims it responsible for a half percent of the contributions to the Linux kernel (0.5%). In reality, Cisco has been a near non-entity as an open source contributor. I'm hoping it's up-and-coming Android-based tablet, Cius, could inspire a change of heart because it's few contributions in its core area of networking have been tantalizing. There's so much more it could do.

I'm not even talking about an open source operating system for its routers -- OS- IOS if you will. Maybe another day I'll argue the merits of that. Cisco keeps its operating system under lock and key (not even allowing it to be licensed out to run on virtual machines for students enrolled in its many certification programs).

OK. Google does the same thing with its core intellectual property. For all of Google's contributions to open source -- too numerous to name -- Google doesn't open source its secret sauce, its precious search algorithms. Quite the opposite, it patents its technologies in that area.

Cisco could follow Google's lead ... driving massive contributions and innovation in the areas that surround networking without opening its primary intellectual property. So far Cisco's attempts have been interesting, but feeble given the size of the company and the engineering talent it commands.

I began looking into them after I saw a blog post by Drew Robb on wondering, "Will Cisco Be the Next to Fall to Open Source?" He argues that Cisco is at risk because of Vyatta. I found his argument mostly ridiculous. Vyatta is a cool product and will likely to do very well. But it's just a kick in the shin to Cisco. Then I noticed that a reader of that blog commented that Cisco was a major contributor to Linux and open source.

Is that so?

As I mentioned, it's main claim to open source fame is contributions to the kernel. One Cisco engineer, Roland Dreier, has been the driving force. He's contributed InfiniBand drivers to the kernel. InfiniBand evolved into Remote Direct Memory Access (RDMA) and this has also been contributed to the kernel.

In order for anyone to use RDMA, with Linux or not, they must have NIC cards that support iWARP. If there are iWARP options other than Intel's 10G NIC cards, I couldn't find them. In other words, the technology is interesting for certain HPC environments, which is an important area Linux serves, but it's not a hot technology elsewhere.

So Cisco has this experienced Linux contributor ... can he help encourage the company to do more with the kernel? Linux supports a wide variety of networking components. Cisco experience could be beneficial, could it not? (By the way, I contacted Dreier to ask his thoughts and am awaiting his input. In the meantime, his blog is an interesting look at InfiniBand.)

Another tantalizing project that Cisco was involved in was the creation of uIPv6 back in 2008. This is a super small open-source, IPv6-ready protocol stack. This stack could be put onto tiny devices and help create the "Internet of things." It is used primarily by Contiki, an operating system for embedded devices.

I readily acknowledge that in July, Cisco also open sourced the Telepresence Interoperability Protocol, which allows others to create third-part add-ons for Telepresence. And Cisco has a few Linux-friendly apps. Packet Tracer is a network simulation program that supports Linux as does Cisco Unified Communications Manager 6.0.

Using isn't the same as contributing. Often if Cisco uses an open source project, it tends to give money, which is good, but doesn't always do that.

For instance, Cisco touts its use of Jabber/Extensible Messaging and Presence Protocol (XMPP). It inherited XMPP when it bought Webex. Cisco is one of the primary financial sponsors of the project. On the other hand, Cisco's IronPort uses FreeBSD. While the company vaguely alludes to contributing hardware and open source code to third-parties associated with this spam-fighting device, Cisco is not named as a FreeBSD sponsor.

Now, please take a moment to wander onto Sourceforge, GitHub or even Google Code and search on the term Cisco. You'll find any number of interesting projects for Cisco products or networking services. Maybe some of these products are contributed by Cisco employees, but after checking a random sample, I didn't discover any. And Cisco doesn't lay official claim to them -- perhaps for fear of having to support them?

For example, one of the most popular Cisco tools on Sourceforge is the Cisco IP Phone Inventory Tool, posted by Vince Loschiavo. He's a former independent Cisco consultant who currently works for Juniper Networks. (Nice tool, Vince.)

Open source is the next iteration of open standards. And open standards is the basis upon which Cisco became an industry giant. An enormous opportunity exists for Cisco to become the Google of the open source networking world. Why leave it to others?

Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.
Must read: 10 new UI features coming to Windows 10