Baby Got Bots

Author: Rich Bocchinfuso – Senior Practice Director

Oh, my, God FusionStormers, look at that bot
It is so efficient it looks like something
One of those DevOps customers would use
But, ya know, who understands those DevOps folks?
They only talk to us, because,
We are the low-cost high-volume box pusher, ‘kay?
I mean, that bot is just so damn confusing
I can’t believe it’s just not a box, it’s like out there
I mean what a waste of time, look
It’s just so, nerdy

I like big bots, and I can not lie
You other FusionStormers can’t deny
That box pushin is deliverin itty bitty value
And when a bot walks in
Your value gets sprung…

My anaconda don’t want none
Unless you’ve got bots, hon
You can do cli or point-and-click
But please don’t lose that bot
Some brothers want to play that hard role
And tell you that the bot ain’t gol’
So they point-and-click it
And I pull up quick to automate it
So FusionStormers, if the bot is sound,
And you want a triple value throw down
Dial 1-900-FIXALOT
And kick them nasty low-value box pushin deals
Baby got bots
Baby got bots

Note:  Of course Anaconda in the context of this lyrical masterpiece is referencing my preferred Python distribution, Anaconda.

When Sir Mix-a-Lot released Baby Got Back in 1992, box pushing was where it was at. I remember that SPARCstation IPX I had on my desk in college, I was the cool kid! Funny how things changed in that two years between high school and my junior year of college.

Back in 1992 I did pretty much everything in Emacs and formatted all my papers in LaTeX. Most of my coding was in C, Fortran, Cobol, Assembler and emerging PC languages like Turbo C++ and Delphi, I did most my scripting in either ksh or csh, made lots of use of sed and awk and was starting to adopt Perl. My window manager of choice was FVWM, today I think how wrong it was that I had to compile my desktop. Just about everything required heavy lifting. 1992 doesn’t feel that long ago to me, but today I code in whatever interpreted language suits my needs, Python, JavaScript, etc… I do most of my coding in the cloud on Cloud9 or in Visual Studio Code, I even have a Python compiler on my Android tablet (Pydroid is the bomb, for anyone interested). I write almost exclusively using Grammarly, I use BibMe to document sources; it’s a totally different world.

This evolution hasn’t been much different for the enterprise. From 1993 until about 2010, my love affair with Open Source had me repeatedly trying to run Linux on my desktop. Today, I don’t fight the machine, I just want what works and what allows me to get the job done efficiently. Gone are the days of a marathon 24-hour Slackware installs spent swapping 20+ 3.5 floppies; those days build character, but I don’t miss them, compiling the kernel daily was a way of life back then… What a waste of time!

Like me, people want to focus on what matters; I always thought understanding the APA and MLA formatting for bibliographies was a waste of time… Not so much that the knowledge was a waste, but the idea that memorizing the format had some value to me, rather than it being something that I referenced when needed. Apparently, I was not alone, because now there is an app for that. Routine and repetitive tasks are meant to be automated; it’s the reason the computer is such a popular device. ‘Routine’ or ‘repetitive’ doesn’t mean that the tasks are the same, as the tasks can be nuanced. This is why we have constructs like variables, conditionals, and loops.

My conversations these days tend to be more focused on process and operational optimization. What makes these conversations interesting is that I can very quickly determine if the customer is looking for a screwdriver turner, or if they see value in a vision that improves QOS by better operationalizing their environment.

We live in a world which is quickly becoming dominated by the Twelve-Factor Apps as the path to increased agility and elasticity.  This transition is being aided by IaaS, PaaS, SaaS, FaaS all orchestrated by robust APIs enabling composable infrastructure which caters to the Polyglot DevOps (emphasis on Dev) engineer.  These concepts are readily understood by those who have already adopted microservices, containerization, infrastructure-as-code as a means to design and build cloud-native applications, but what about the legacy infrastructure and monolithic application?  I am trying to think of one. Hmmmm… Maybe Email, oh wait, that’s a SaaS play. What about the traditional RDBMS, wait, that’s a PaaS play. What about ERP, MRP, CRM, HRM, ITSM, ITOM, etc…, a SaaS play again. So the idea of orchestrating and automating that which we consume via API is pretty straightforward, that which we call the “cloud” is just composable infrastructure.  Maybe we own the infrastructure, aka “Private Cloud”, maybe we consume the infrastructure or service owned by someone else, aka “Public Cloud” or maybe we do a little of both, aka “Hybrid Cloud”. Today we see the emergence of Metal-as-a-Service, where we can begin to apply orchestration and automation to the infrastructure hosting legacy applications. Legacy applications are often always-on, lacking the agile and elastic properties of cloud-native apps, but with Metal-as-a-Service we can begin to apply the principles of Site Reliability Engineering at a lower-level allowing us to orchestrate, automate and improve resiliency without the need to re-platform applications.  The movement toward composable infrastructure (aka infrastructure-as code) has provided the ability to programmatically automate processes, enable self-healing, dramatically improving Mean-time-to-Ack, Mean-time-to-Engagement, Mean-time-to-Workaround, and Mean-time-to-Resolution. When we take all these APIs, automation and orchestration, add in ChatOps and bots we end up with a platform which increases the velocity with which we can interact with infrastructure, workloads, services, and applications, improving our visibility and communication, and accelerating our engagement while reducing the risk of human error. Bots automate tasks, apply intelligence and increase our velocity while reducing the risk of human error, is there any question why “My anaconda don’t want none – Unless you’ve got bots, hon”.

FusionStorm is helping customers at all stages of the lifecycle, from legacy to converged, to hyper-converged to composable infrastructure by applying best practices, bridging the skills gap and changing the way they operate.  The transition from reactive point-and-click infrastructure operations to automated infrastructure operations requires a shift in thinking, away from architecting for uptime to architecting for failure. Good news, robust APIs and programmatic interfaces exist almost everywhere in modern infrastructure allowing for the application of practices which improve agility, velocity, reliability, and quality.  Business outcomes such as faster time-to-market and reduced operational risk can now be realized by nearly every organization. This pivot requires a different set of skills; not every organization has an army of polyglot developers waiting to supplant the traditional system administrator, not every organization has system administrators with the desire to become developers, this is where FusionStorm can help.