Why the Network Industry has been Stuck in the 1980s:  Cisco's Embrace of Complexity

Why the Network Industry has been Stuck in the 1980s: Cisco's Embrace of Complexity

Every time a major corporation or government entity makes a large purchase of Cisco Systems network routers and switches, they are casting a vote for the perpetuation of 1980s-class technology. 

In the past 18 months, I’ve migrated my personal IT environment to a MacBook, an iPhone, and an iPad (away from Windows and Android OSs). Why? The usual reasons – better reliability and security, a superior user experience, near seamless interoperability, and they “just work.” I really like my Apple products. No wonder Apple has been the most important technology company, by a long shot, for several years. 

So why can’t network products (routers, switches, etc.) become … more secure and reliable, offer a superior user experience, interoperate near seamlessly, and “just work”? Seriously.

Ethernet and IP networking is embarrassingly complex, unreliable, arcane, and parochial. That results in very high operational costs, poor security/high vulnerability, and nothing close to five nines reliability. In almost any other product category this would be considered unacceptable. Network technology has changed very little since the late 1980s, with the exception of faster speeds/feeds and some additional protocols and features. 

It is not coincidental that the networking industry is dominated by an 800 pound gorilla, Cisco Systems. They have held greater than 60% worldwide market share for Ethernet and IP products for a very long time. In some segments, Cisco’s market share exceeds 80%. No other router and switch vendor gets beyond single digit market share, and most are in the 3% - 4% range.

Cisco’s dominance in Ethernet and IP networking is the reason the networking industry delivers a 1980’s experience (think DOS PCs if you are old enough) – command line interfaces, extensive manual configuration, trial and error, failures and outages, difficult troubleshooting, etc.

So how did they get 60% + market share? Cisco’s networking products do not do anything different, with a few minor exceptions for relatively unimportant proprietary features, than other network vendors. Their products are based on Ethernet and Internet Protocol industry standards, use many of the same merchant chip sets as their competitors, and offer similar performance characteristics. They are usually not best-in-breed in any sub-category. Cisco builds quality products, which sometimes have issues, just like many other companies.   

Obviously part of the answer is brilliant business execution, dating back to the early days in the late 1980s, lead by John Morgridge and taken to the next several levels by John Chambers. They deserve a lot of credit. (Let me also add – this post is not an attack on Cisco, a company whose business I have long respected and admired. Rather, it is an explanation of how a key technology segment got to where it is today).

The real answer is something else: the perpetuation of complexity, and the leveraging of that complexity as a business advantage. 

So when complexity is a strong business advantage, what is the incentive to change, to develop technologies and products and standards that make networks less complex, more reliable, and less expensive to operate? Answer: there is none. Cisco has no incentive (in any serious way) to innovate and make the technology inherently more automated, standardized, and easier to operate. Their market dominance is dependent on complexity. 

With complexity comes fear, and the only perceived safe haven for that fear (FUD) is a heavily dominant market leader, who seemingly provides the basis of a reliable outcome. 

In addition, Cisco dominates the market in IP networking training, education, certifications, and the like. Most network engineers have achieved Cisco certifications. Cisco trained engineers will be heavily biased towards Cisco products. Furthermore, Cisco trained engineers embrace the complexity, for the obvious reason that their skills become special, valuable and well compensated. These people do not want networking to become easier and more reliable. Cisco has created the perception – and it is only a perception – that Cisco trained people can only work on Cisco equipment. Industry and government are complicit in buying into these erroneous assumptions. This influence on the labor supply chain drives costs up greatly. The reality is that a good engineer, trained in IP and other network technology, can work on any quality product that complies with standards and is well designed. 

It can (and should) be argued that this is a free market system. Why don’t other companies improve the technology if this is such a big issue? Why isn’t there a plug-and-play network ecosystem? Where are the VCs, the entrepreneurs and other opportunists? It is in fact starting, on a few different fronts. 

  1. Fabric technology, for both the Ethernet and IP layers, substantially simplifies networking. By placing a layer of automation and simplification in front of standard network protocols and operating systems (somewhat like placing a GUI in front of the UNIX OS), most of the manual configuration and cumbersome architectural gymnastics are eliminated, providing a much bigger step towards “plug-and-play.”
  2. Software Defined Networking (SDN), with network intelligence and control centralized for automation, advanced control, and integration of the network with business applications, will provide a layer of abstraction above network hardware.   As SDN technology matures with time and investment, resulting automation will lower OpEx, improve operations, and enable new network services and capabilities. The OpenFlow protocol, a subset of SDN, is the first true alternative to the Internet Protocol in a long time.
  3. Operational automation tools. Beyond scripting, there is a new generation of tools to help automate the network, leveraging products such as Puppet and Chef as well as new middleware and data bus products.

So Mr. or Ms. CIO/VP of IT/etc., when you place that next big order for Cisco products, do you understand that you are helping to perpetuate unacceptably high complexity and costs and the continuation of stale technology? Cisco has no incentive to make anything more than proprietary incremental improvements. They have every incentive to have your network continue to be overly complex and dependent on only them. There are some things that you can do, however. 

  1. Give a percentage of your network spend to an alternative to Cisco. Those are the companies that are most vested in changing the status quo and are investing in new technologies and driving new open standards to lower overall costs. Two vendors provide better leverage to lower CapEx. This will also train your staff to be less dependent on a semi-monopolist, with all the premium costs associated with that captivity.
  2. Test an Ethernet or IP fabric solution and gauge the potential OpEx savings.
  3. Start a serious look at SDN and work with a company that can deliver expertise in this area. Start planning for an SDN trial – it is coming – and conduct a proof of concept when you are ready.
  4. Insist on engineers that are network professionals, not Cisco clones. A true professional is all about the protocols and technology and operational excellence, not about a specific vendor’s products.
  5. Push for training in schools, the military, and other organizations that is about the technology, not about one vendor. When a school accepts and uses Cisco’s course material and curriculum, they are indirectly subsidizing the high costs and continued complexity of networks. Industry, the schools, the military, and other vendors must develop and/or deploy technology centric, not vendor-centric, training. 

The networking industry has just started to be stirred from a long period of complacency. Now is the time to move beyond the 1980s and start towards a modern network world. Every serious technology executive can contribute to the advancement of the state-of-the-art, and the first step is to not automatically give that next big order to Cisco.

 

 

 

 

 

Sanna Yang

APJC Incubation sales -Success Track/CX cloud

6y

With the DNA, intuitive network approaching, Cisco break down the barriers it used to embrace. Looking forward to see the industry got stirred up.

Like
Reply
Joe Howard

Working to Digitally Modernize the Army - Edge to Hybrid Cloud

6y

In early 2016 I posted this article on why I thought the network industry was a relative laggard in terms of progress. My hypothesis was (and still is) that the inherent complexity of IP networking allowed one company to skillfully exploit risk and uncertainty to dominate the industry. Why drive serious innovation and technology simplification when complexity is a business advantage? This insight stirred a lot of healthy debate which I hope will continue. Fast forward to today, and we see change in network technology finally starting to take hold. SD-WAN, which offers simplification and reduced costs of WAN links, will be a $1B market in a few years. While Cisco recently made an acquisition in this space and will surely do well, there are over a dozen start-ups in this area and it is unlikely that Cisco will dominate the WAN going forward as they have in the past. VMware is dominating data center network virtualization with their NSX SDN platform, apparently justifying their $1.2B acquisition of Nicira. OpenFlow is becoming standard on most switches and routers, enabling at least some applications to leverage the centralized intelligence and automation of an SDN controller. Networking customers seem to be recognizing that in the era of SDN there will be an "open" path and a proprietary path, and open is usually preferable, especially as the technology matures. There are other examples as well. So the network industry is beginning to seriously change, and Cisco will not dominate in the future as they have in the past. Indeed, they are losing market share in their L2 and L3 switch and router market. Weep not for Cisco, of course, because their diversification and software focus is making them a different company, which I expect to be very successful in their own right, but not necessarily as a network infrastructure company. The network industry needs to continue to push for innovation and virtualization. It is taking a lot longer than everybody thought it would, but Network Industry 2.0 appears to be ascendant.

Mark Zimmerman

Systems Engineering at Palo Alto Networks Stategic MidAtlantic

7y

Excellent ! I see lots of good debate below but in summary, Cisco tends to overly complicate technology. We use to joke at Cisco that we had to justify all the PhDs for all the complexity. ( I really didn't know many PhD's are employed at Cisco) My take is intelligent people should take something complex and simplify it. Cisco is better at marketing then innovation, if Cisco could swap that around then maybe we could see a change but as your write up mentioned, what would compel Cisco to do this ?

Jim Barker

Constantly learning new technologies in an ever changing technical landscape.

7y

Thank you Joe. For a slightly different perspective, a review of a Cisco acquired company called Meraki is in order I think. It is a networking company using cloud configuration and maintenance. We deployed our entire wireless infrastructure with Meraki WAPs managed entirely in the cloud. My experience with it is a very positive one and while I appreciate the complexity of networks, I specifically chose Meraki because of it's ease of deployment and maintenance. They also have cloud managed firewalls and switches.

To view or add a comment, sign in

Insights from the community

Explore topics