People Time versus Product Costs
#devops #SDN #SDAS #Cloud Every recent major innovative change in networking has ultimately been about operational efficiencies; about people time and not product costs.
Financially, IT is a study in capital versus operating expenditures. It's a study in short term versus long term; in product cost versus people time.
The major innovative shifts in the industry - starting with cloud and moving on to SDN and devops - have really been about the latter - operations - despite the initial laser-focus on capital cost savings. That's because over time, the bulk of the cost of any product, solution or service is always in people time. Whether measured in dollars or hours, most of the costs in IT are the result of configuration, management, upgrades and maintenance over time. Refresh cycles are generally the demesne of financial analysts and vendors, but they play heavily into the reasons why operational costs are almost always higher than capital costs. IT simply doesn't replace its entire network or any large portion of its infrastructure often. There are very predictable cycles on which such events occur and in between, the operational costs are wracking up the bulk of the total cost of ownership.
Battle of Standardization versus Commoditization
The cost of an OpenFlow-enabled white box switch, for example, is certainly lower than its proprietary counterparts from well-known networking leaders due to commoditization. But it's still only a small portion of the total cost of ownership, which needs to take into consideration the operational savings from greater efficiencies driven from standardization that enable automation and orchestration to reduce people time-related costs.
Operational costs end up being the great equalizer; the people time saved by automating and orchestration the network is true whether the target of that automation and orchestration is a white box or a top-of-the-line proprietary switch.
It could be argued, in fact, that proprietary hardware, having a long and proven history and existing partnerships (ecosystems) will ultimately net even greater savings because of the ability to leverage those existing partnerships to integrate across multiple systems. That means less people time spent by IT integrating and testing and tweaking, which in turn drives down the operational costs associated with the solution.
Regardless, the reality remains that it is operational efficiencies and thus people that are the ultimate beneficiaries of recent transformational technologies. The reason it may be acceptable to "rip and replace" that enterprise-class network fabric with an OpenFlow-enabled, white box switching fabric is precisely because the long term return on investment is so compelling. But that's true of proprietary, more expensive solutions, too, that come with guarantees that can't be offered by the disposable-driven nature of commoditized products.
Craig Matsumoto summed it up nicely in his article, "White Box Week: How Cheap Is ‘Commodity?’"
Even proponents agree that capex — the cheapness of the boxes themselves — won’t drive this market. Capex gets customers’ attention (who doesn’t like the idea of spending less money?) but it’s really the operational savings that hook them.
In the battle of capex versus opex, opex is really the driving force behind almost all the "trends" in networking and infrastructure. Cloud? It's about operational efficiencies and service velocity. Devops? it's about operationalizing continuous deployment and delivery. SDN? It's about network service velocity, and operationalizing the network.
It's all about people time, not product cost.