One may not be able to teach an old dog new tricks, and networks might not adapt well, either. A handful of networking experts discussed their thoughts on networking innovations, including how traditional thinking can delay networks from moving forward and if change is possible.
Networking expert Greg Ferro listed on his Ethereal Mind blog four networking innovations and mindsets that engineers should re-evaluate. He covered misunderstandings about the general flow of networks and network protocols to newer architectures, such as software-defined networking (SDN).
These issues highlighted an overall disconnect among the technology, engineers and users. According to Ferro, the main pain points of networking include the following:
- Network engineers don't think of networks as end-to-end, but as hop-by-hop, which weighs specific hops and individual devices more heavily than network flow.
- A network is not a singular entity, but a combination of many interconnected systems.
- The idea of self-configuring systems is all hype, and the systems themselves are unpredictable and use weak protocols.
- SDN is still all bark and no bite. It lacks standards, so it's unclear how connectivity and coordination among various controllers will work.
"My concern is connectivity and coordination between these SDN-controlled networks, because there are zero standards on interoperability between these," Ferro wrote. "None of the existing standards apply to these new systems."
By reconsidering these four areas, engineers might be able to drive some necessary innovations in networking. Embracing change in these areas could help move networks forward and out of the past.
Read more of Ferro's thoughts on necessary network changes.
Does traditional thinking thwart technological transformations?
Enterprise networking analyst Steven Schuchart Jr. tackled similar networking innovation topics on Global Data's blog. For example, he said traditional networking thinking may prohibit the industry from advancing further -- campus networks, in particular.
Campus networks need agility, security and efficiency. But vendors and engineers think of campus networks as divided wired and wireless networks, which hinders development in these areas, Schuchart said. This way of thinking also impedes the campus network from positive change. A holistic view of networks could spark networking innovation, however, and benefit different types of networks.
A holistic network view would not eliminate specialized vendor services, according to Schuchart, but vendors would need to bolster their products and services to emphasize more aspects of the network. And while network teams often have a negative perspective on vendor lock-in, it may be an inevitable circumstance of the industry.
"The needs of the business mean that gaining operational efficiency, agility and security are more important to the organization than IT's ability to switch vendors," Schuchart wrote.
Read more of Schuchart's analysis of old campus networks and their need for new tricks.
Is networking innovation truly transformative?
Tom Nolle, president of CIMI Corp., also discussed network transformation in a recent blog post. In particular, he said the most fundamental way to transform a network would be to build networks differently. Not only would organizations change the way they build network architecture, but vendors would need to adapt, too.
The main reason why previous attempts at networking innovation failed is because convincing network teams to change ways is difficult -- especially if the change involves a new, unproved and potentially unreliable architecture. Also, the last innovation that drove transformational change was the creation of the internet, and a disruption like that is unlikely to happen again.
However, network transformation could be possible by separating service and transport networks and improving Opex to link transformation with new revenue, Nolle said. Virtualization could play a significant role in separating service and transport, as it could be a different way to construct these networks with virtual tunnels to route traffic.
"I firmly believe that a two-tier transport/service model would work better," Nolle wrote. "Vendors could define a migration path to it that would satisfy operators' cost and benefit requirements and even provide a fairly safe transition strategy -- one that you could fall back from early on if needed".
Dive further into Nolle's thoughts on network transformation.