As a Core Solutions Architect here at Broadcom, I am on the VMware Cloud Foundation (VCF) front lines. I am the technical lead on discussions with customers about VCF, their current use cases for VMware, and what it all means for them.
The challenge, invariably, is that even fans of VMware continue to pigeon-hole VMware software as solely “that hypervisor,” and I’m hoping to broaden your horizons here and make the case for VCF as a private cloud platform. I feel compelled to point out that VMware software has been part of a software-defined private cloud platform for the better part of a decade, but that’s another blog post altogether.
This post is mostly for vSphere engineers, but if you have engineers who report to you, I think you will get a lot from this post as well.
First, some nerdy history.
Conway’s Law
Conway’s Law states that
“[O]rganizations which design systems . . . are constrained to produce designs which are copies of the communication structures of these organizations.”
– Melvin E. Conway
In other words, your IT infrastructure is only as good (or as bad) as your business communications structures. Let’s call them “silos”.
If you apply this to how x86-based data centers have been done for, let’s say 30 years (give or take), IT infrastructure becomes siloed, which has proven to be horribly inefficient: IT infrastructure silos are rife with internecine politics, job protectionism, territorial protectionism, lack of trust, lack of communication, finger-pointing, and endless lists of human-driven tickets.
This leads to the crux of the problem: barriers to execution.
The traditional siloed approach in our data centers, as above, is not just inefficient: it has failed.
VCF Leading the Private Cloud Platform Paradigm
If you are an infrastructure engineer, have you ever thought about why people might migrate workloads out from under you? The likely answer is because users and consumers of your infrastructure found a way to do it faster elsewhere. IT end-users (of whatever type) expect things at their fingertips, or delivery of resources in seconds or minutes; not days or weeks or months.
But the paradigm shift is that with VCF, you become a cloud admin. You administer a private cloud platform now. This means that everything is software defined and your priorities change:
As a cloud admin (as opposed to the limiting “VI Admin”), you are no longer tied down with the toil of manual life cycle management or hardware configuration. You aren’t supposed to care about what type of physical switches run on the network. That gets obfuscated away so you can focus on what’s really important: getting what your users need, and getting them what they need through self service in seconds or in minutes rather than in days or in weeks.
What Does This All Mean for Engineers on Day 2?
It means that your life is defined by “Policy-based management“. You have policies that you define for just about everything: VM sizing policies, storage policies, network policies, security policies, and so on.
Now, when you provide self-service to your consumers, or if you provision the infrastructure yourself, you define the policies that apply and the platform does the rest.
No more, “We’re waiting for the network team to create firewall rules.” No more, “We’re waiting for the security team to approve the VM.” No more, “We’re waiting on the storage team to zone the LUNs.”
You are now proactive not reactive in everything you do.
Have No Fear, Guardrails and Automation are Here
In my experiences, VCF as a private cloud platform causes two fears from engineers, both of which I would like to dispel here:
- If I am going to provide self-service, then what about if humans make a mistake and cause an issue?
The simple answer is that you control the guardrails for self service. Afraid someone will roll out 40 VMs instead of 4? You control that through guardrails and limits. Afraid someone is going to put a VM on the wrong network? Your policies define and limit who can do what, when, and where. - Doesn’t this mean more work for me?
In the long run, no. Your work is front-loaded now rather than ad-hoc. One of the (ironic!) main objections I get from the engineers with whom I speak is that they have no time for new projects. The whole point of the automation built into VCF is to give you more time to “do the needful”. Your efforts are now less on “Keeping the Lights On” (KLO), and more on innovation for your organization.
How Not to Approach VCF
The mistake I see people make is viewing VCF as an a la carte bundle. It is not.
Yes, you can ramp up gradually to VCF by implementing the product’s various compute, networking, storage and management capabilities (we refer to this as the “Crawl, Walk, Run” approach).
But the most value is realized by implementing VCF as a full stack, so we recommend a roadmap that gets there to gain the most value.
IT Infrastructure as a Force for Innovation
Time-to-delivery is important for every organization with which I speak. So, how would you like to be viewed as an individual contributor who actually makes money for the business, rather than as nothing more than an expensive cost center?
Also, the skills you gain by becoming Cloud Admin through your deployment and operation of VCF are also keeping you up to date with the industry.
VCF is literally good for your career.
