The first car I got decades ago was a simple mechanical beast – you’d push something, and a cable would make sure something else moved somewhere. I could also fix 80% of the problems, and people who were willing to change spark plugs and similar stuff could get to 90+%.
Today the cars are distributed computer systems that nobody can fix once they get a quirk that is not discoverable with level-1 diagnostic tools. Trust me – I’ve been there, and all the vendor experts did was ever-more-ridiculous part swapping, ending with “let’s swap the main CPU because we have no other idea and sometimes it helps”.
An old mechanic who was willing to look at individual measurement points and diagrams to figure out what the diagnostic tool was missing eventually fixed the problem. Turned out it was a leaky pipe – a problem never considered by experts creating the diagnostic tool.
As the pundits love to tell us, we have to make networks as simple as driving cars. The single-image systems (from QFabric to ACI), SDN, and intent-driven networks should make networking a commoditized simple-to-use system.
They might eventually get there, but there’s the pesky law of leaky abstractions. Once you hit it, and your network is down, what will you do? Will you be able to fix the problem, or will you wait for $vendor technicians to twiddle with the diagnostic tool, and finally for the $vendor experts to scratch their heads?
What will happen to your business in the meantime? Can you survive without a network? Can you rent a network like you could rent a car when yours is under repair? Can you reformat the network if you mess it up like you could an infected laptop?
For some environments, the answer is YES – or at least the cost of having someone competent on staff or on call is higher than the cost of occasional downtime. Are you one of them? If not, carefully consider how you plan to deal with the magic new technology once it breaks, and (paraphrasing Gartner) start investing into premium people instead of premium vendors.