Cisco Renews Focus on Network Engineers

As Cisco expanded beyond switching and routing, network engineers had to put up with buggy code and a loss of innovation. The company is finally showing network engineers they haven’t been forgotten.

Ethan Banks

January 14, 2016

5 Min Read
Network Computing logo

To manage a network is to love it. Is that too sentimental to describe hunks of electrified metal and silicon that happen to bolt nicely into a 19-inch rack? Perhaps for some, but not me. I and many other packet pushers have a passion for networking that drives our work, motivates research, and inspires discussion.

Most networking engineers have a relationship--good or bad--with Cisco. Cisco started in the routing business, then expanded into switches, and over time added firewalls and IPS devices (to mixed reviews). Voice over IP became a popular choice to replace aging PBX and key systems, and Cisco took a strong position.

All of these plays made sense to those of us working on Cisco gear. Cisco, mostly through acquisition, was entering markets that played to its strengths and allowed it to walk into sales opportunities with a complete, integrated networking package. Network engineers worked on this gear, formed their relationships and, in many cases, became internally aligned with Cisco.

Then it got weird. As stockholders looked for continued growth, Cisco had to cast its gaze ever-wider to acquire companies that could boost the bottom line. "Linksys? That's consumer networking gear. Ew. But I'm sure it will be fine," engineers thought. "Scientific-Atlanta? They make set top boxes? What the heck am I going to with that? But I'm sure it will be fine." "Pure Digital? As in those little Flip cameras? I don't I get it, but I'm sure it will be fine."

[Cisco is gearing up to expand its bottom line via services. Find out how in “Cisco Aims to Increase Services Business.”]

But Cisco wasn't fine. Cisco spread itself thin as it chased markets outside its networking core. The perception among network engineers was that Cisco was shifting away from its fundamentals--creating value-added networks--and into a mode of throwing everything at the wall to see what might stick.

That had repercussions on our own networks. First was software quality. While bugs are common in any software product, the sheer volume of bugs showing up in networking software--even in mission-critical code for core equipment--forced engineers to choose between bad and ugly: Stick with the devil you know, or trade in the bugs you have for new ones.

The other issue was the slow release of new products and features, which prevented Cisco from leading in certain market segments. One need only look at the Nexus switching line, formidable though it is, and compare it to Arista 7500E's high port density and aggressive per-port pricing, or to Brocade VCS's ease of use to see where vendors leaped past Cisco on performance or technical innovation.

Another example is in network operating system virtualization, such that network engineers can model their networks to test changes and learn new features in a low-cost, low-risk environment. While the open-source (and arguably rogue) Dynamips and GNS3 filled this gap for Cisco engineers, Juniper released Junosphere and Arista debuted vEOS, but Cisco had nothing official available.

Next page: Change You Can Believe In

But I believe Cisco’s focus is shifting back to the customer. And not just any customer, but the ones that makes up its core business: those who run value-added networks. This change has been building for a while. For instance, Flip is dead and Linksys has gone to Belkin. While Scientific-Atlanta is still on the books due to Cisco's interest in the global video market, rumors that Cisco considered selling it are a Google search away. More substantial changes appeared during Cisco Live in Orlando, where I felt like my old Cisco was back.

Here’s why.

The first customer-focused announcement was that of the Nexus 7700, which boasts front-to-back airflow, along with the F3 line card announcement. True front-to-back airflow is a big deal in big data center design. The 7700 is also providing a way forward for shops needing to aggregate lots of 40-GbE and 100-GbE ports in a single chassis. Meanwhile, the F3's unique ASIC hearkens to the customer need for a single line card that can do it all.

The next big piece of news was that of Nexus Validation Testing. NVT is a new software testing program within Cisco designed to improve quality assurance. The NVT process builds network topologies like a customer would build them, configures software features like a customer would configure them and then tests code. Failures in the testing process prevent the code from being released to the public.

If you're wondering how Cisco did testing before, my understanding is that features were tested individually, and not in the context of a fully built network infrastructure with several features deployed simultaneously. While the old process could detect specific bugs in individual features, the new process can detect the sorts of bugs that only appear in a complex production network environment. The result is that NX-OS (and eventually other operating systems in Cisco's lineup) will have been thoroughly tested in customer-like environments and the bugs resolved before being made available to customers. No longer will customer networks be doing Cisco's QA work for dot-zero releases.

Cisco also demonstrated its Virtual Internet Routing Lab. While not officially released yet, VIRL lets networking teams mock up a network topology and run data through it to validate designs, test features, and so on. My understanding is that it will come in three flavors: cloud-based, appliance-based and stand-alone installation (on a system with lots and lots of RAM). VIRL supports models that run various flavors of IOS, as well as NX-OS. And while strictly a rumor, I've heard from several different folks that a free version aimed at engineers working on certifications is likely.

Detractors will point out that perhaps front-to-back airflow should have showed up with the first iteration of the Nexus 7000 line. And perhaps mocking up customer networks to validate new network operating system code is an obvious QA step to take before releasing the code to customers. And maybe VIRL has been a long time coming. Those are reasonable criticisms. But my point here isn't to beat up Cisco for early design decisions or for how long it's taken for NVT and VIRL to become realities. Rather, I'm elated to be able to say that Cisco is heading in the right direction for its customers. Yes, it's a big ship to turn, but Cisco is listening to feedback and making the changes needed.

 

About the Author(s)

Ethan Banks

Senior Network ArchitectEthan Banks, CCIE #20655, is a hands-on networking practitioner who has designed, built and maintained networks for higher education, state government, financial institutions, and technology corporations. Ethan is also a host of the Packet Pushers Podcast. The technical program covers practical network design, as well as cutting edge topics like virtualization, OpenFlow, software defined networking, and overlay protocols. The podcast has more than one million unique downloads, and today reaches a global audience of more than 10,000 listeners. Also a writer, Ethan covers network engineering and the networking industry for a variety of IT publications and is editor for the independent community of bloggers at PacketPushers.net.

SUBSCRIBE TO OUR NEWSLETTER
Stay informed! Sign up to get expert advice and insight delivered direct to your inbox

You May Also Like


More Insights