Thank you to all who generously shared experiences and views on my “Weave An Automation Fabric For All Technology — Compute And Beyond” blog. In that blog, I emphasized application developers’ need for holistic and integrated automation capabilities. Here, I’m stressing why today’s tech operations can’t stand without automation.
Consider that you operate a live streaming business such as Hotstar. You are broadcasting important news, a globally watched sports event, or a popular reality show. The event broadcast runs for an hour or two. Within that short duration, it is not atypical to experience the number of live users spiking from hundreds to millions before plummeting back down.
Variance Feeds Challenges
As the user load grows or shrinks, all resources including storage, compute, network, security, and application instances must flex. As a tech executive, you have a few challenges, including:
(A) You can’t afford to deploy hefty infrastructure in anticipation of that erratic viewership.
(B) You can’t afford to take an hour to adjust.
(C) Manual scaling of resources is an operational nightmare, unrealistic, and undermines your customers’ trust in you.
Dramatic instantaneous flexing, like what Hotstar needs, calls for real-time resource scaling in response to changing user traffic. Your ability to implement real-time execution is underpinned by a clear understanding of several key performance indicators. First, monitor real-time experiences and user load trend. Second, understand resource behavior. Third, anticipate possible results of your resource flexing actions. Nothing stops you from executing these as independent, unintegrated, stand-alone tasks. The challenge, however, is that the longer you take to respond, the higher it costs your business.
Automated Tech Operations Is A Panacea
You need observation (not observability) tools to analyze every KPI, as well as perfected, automated playbooks. Each automated playbook corresponds to a likely, impacting event. As you learn through various experiences and situations, you end up growing your library of automated closed-loop chains of actions.
If the above vision seems trivial, consider that you must deliver dependable services that have a sphere of security that also expands or shrinks. For this, you must apply the right security profile to newly provisioned infrastructure.
How About Naysayers?
I would not be surprised if a few argue that their applications do not need to be as dependable or flexible as Hotstar’s. I recommend internalizing the promises your company makes to your customers, employees, and other stakeholders. You’ll realize that today’s enterprise applications are not drastically different. Your operating scale may vary, but your clients’ expectations do not. The question you should ask yourself isn’t whether or not you want to automate your tech operations. Challenge yourself. “What does it take for me to build a dependable automated stack?” Your customers expect you to deliver dependable services.
As always, I welcome your thoughts, experiences, opinions, and comments. For a conversation, please schedule a Forrester inquiry call. I look forward to speaking with you.