Author: Chip Ernst
Edge is bringing us back to the days of distributed computing. For so long we’ve had hard choices to make between hands-on control and the promised economies when we let go.
The middle ground in between was always a compromise, neither as fast or reliant as our own racks, nor as flexible and manageable as when disintermediated to a public cloud. But this presumes that cloud is a platform to which we subscribe, rather than a way we can organize all of our IT systems to operate as.
This is the gift of Kubernetes, the enabler to running our applications and their infrastructure, whether here or there, physical or ‘cloud’, in a uniform and consistent way- it can all be ‘cloud-like’ and bring us the best of both modes. In particular, efficiency & mobility as the key outcomes of this new mode whether that be cost, performance, manageability or all of these that you measure.
The notion of edge and cloud (hybrid computing) isn’t to just suggest some on-premises and some cloud, maybe you’ve already been doing that for several years already. Hybrid is the next step after you’ve already taken that leap into cloud- and come to realize that you need to pull some runtime back to earth for cost, performance and control.
Hybrid is the promise of adding greater programmability to your global information systems so that you can leverage this new on-premises or edge capability like you could manage cloud- with uniformity, by software, and with new characteristics that were always out of reach with legacy on premises solutions. Things have changed after all, containers, microservices, and democratisation (commoditization) of the underlying hardware and cloud services beneath a uniform orchestration layer.
Let’s remember the dream, the X as code revolution: it’s now possible, after all, to procure state of the art compute, network and storage capability off the shelf with hardware capabilities rivaling any major system vendor. Robust and capable software is now readily available whether open source or not for all the requirements that you dictate. This software provides greater efficiencies across the environment, matching or surpassing performance characteristics, simplifying administration and eliminating the obligation to pay servitude to one of few tech giants in the industry.
However, if we were to build our hybrid model with legacy approaches, using established technologies that carry those tired tendencies which set our data in concrete, minimize any possibility of practical unified management, limit us to using ‘prior generation’ virtualisation strategies; we aren’t moving forward, but have kept our comfortable old well worn chains.
By embracing Kubernetes and the honest application of those inherent capabilities, we have new possibilities to explore, greater efficiencies to enjoy. Now we may containerize production databases, orchestrate the edge and on-prem via Kubernetes equally along with cloud workloads. This indeed presumes we have competent software capabilities matching or surpassing the ‘enterprise’ level of features demanded. And solutions do now exist for controlling, securing, storing and moving your data which provides a cloud-native approach to your runtime whether it’s one prem or in the cloud. All this without giving up the richness of features and capabilities the industry has accumulated over many years, it’s no longer a compromise.
It’s only now that one can leverage these cloud-native environments like never before, moving between edge and cloud- managing those disparate environments like one for the business outcomes we all seek: cost management, speed to market, improved control, greater capability. Alas, with bright new horizons ahead, we must still navigate the woods to find the sunlight.
As we’re seeing each new day this frontier unfolds before us, there is some of the overstatement to what some new cloud native technologies may actually yield over yesterday’s tried and true offers. Is all cloud native the same? Clearly not, so during this expansion phase where multi-cloud and hybrid-cloud in their truest forms begin to be deployed with vigor, reminding ourselves of why these architectures are distinguished from their predecessors is important.
Containerizing a monolith, distributing it, and calling it cloud-native is not comparable to a microservices architecture that spans pods, clusters and clouds. If we are trying to attain new levels of efficiency and control, then we best ascertain that we indeed are implementing these new environments within this new paradigm by design.
When we apply well-designed software, leverage the most scalable technical concepts, with innovative control freed of minutia management practices of the past, we can see not just the forest through the trees, but the bright pastures beyond. It’s then that space between the trees becomes the music and we are no longer pulling weeds but enjoying the landscape.