The recent acquisition of VMware by Broadcom has caused ripples of uncertainty. Licensing changes announced by Broadcom have raised concerns about the future of VMware’s pricing structure and its commitment to open-source technologies. As customers consider renewal quotes that in many cases they simply cannot afford, competing solutions have re-tooled their marketing and sales campaigns on the quick in hopes of positioning themselves as THE alternative. Meanwhile, the internet is teaming with emotionally charged reactions from long-time VMware disciples who feel betrayed by their religion. As the clock ticks down on current license agreements, the move from a la carte solutions to bloated bundles just adds fuel to the fire. The result is confusion, frustration, and in many cases good old fashioned analysis paralysis.
The seemingly ‘intellectual’ IT community should ratchet back its emotions. Working from the neck up will, as it’s done in the past, win the day.
A tried-and-true methodology for moving forward exists. Countless solutions have been evaluated based on finding the overlap between business needs, technology, and budget. This simple Venn diagram has been collecting dust. Because everyone loved Pat Gelsinger, and the perpetual licensing he rode in on, we haven’t needed it. But underneath the reddit posts and virtually planted flags testifying that ‘everything is going to the public cloud,’ there are mindful, reasonable IT professionals that actually know how to figure this out. If only they could set aside their albeit trendy, but not very productive Hock Tan rage.
Where does this secret logic hide? How can an entire industry follow Tay Tay’s seemingly simple advice and just ‘shake this off? It’s actually quite simple.
Let the data speak
For a generation, solutions architects, pre-sales folks, self-configuring customers, and anyone with purchase order authorization have said yes to the core wars waged by Intel and AMD. If one socket was good, two was better. Customers and hardware vendors alike maxed themselves out at 32 cores per socket and their only complaint was that it took two weeks to ship. But shoulder pads, neon, and every hair band of the ‘80s showed us that excess always seems reasonable … until it’s not.
Reckoning has occurred. Buying more hardware than we need actually costs more – WAY more if you are a VMware customer.
It’s time to right size our infrastructure
The tools exist. OneIQ, Live Optics, RV tools – pick your poison. After a decade of looking at reports generated from these tools, I can tell you that they all tell the same story. The vast majority of customers are using less than 5% of their CPU capacity. Those same customers are losing their minds about their VMware renewal quotes.
And this is not just about Hock Tan and his band of unwilling VMware evangelists (in fairness, they DO have mortgages to pay). Newsflash: we pay (Pat G and funky-bunch) more for more CPU. Add to that, for a second socket, per Intel / AMD Core Wars guidelines, we must maintain a ‘balanced’ memory configuration.
More precious IT dollars are spent
Also, in fairness, while Hock Tan does his best to crash land a seemingly functional aircraft called VMware, the likes of Michael Dell, and even Pat ‘Switzerland’ Gelsinger are still laughing their way to the perpetual fountain of private equity. But I digress.
The point is that as an industry we have been lured into super-sizing everything. We have been led to believe that because we have been virtualizing, we have been making efficient use of infrastructure, even if we weren’t actually using it. And while Hock Tan and his merry band of ‘it’s coming directly from Hock Tan’ mortgage victims serve as fantastic scape goats, this problem existed WELL before Broadcom acquired VMware. We all fell for it. And just like we blamed all the stupid stuff we did in the 80s on MTV, we want to blame Broadcom.
This is on us
It’s hard. AMD has created a fantastic David and Goliath story. We are all paralyzed by a hyper-scaler induced fear of scale. Certainly, if I have enough cores on prem, I can assuage my anxiety over the fact that I don’t have the budget or talent to support migrating to the public cloud. More cores and more sockets solve for a problem of scale I will never encounter, in this budget cycle or ever.
If you have read this far, you have either fallen victim to the Core Wars that AMD and Intel have forged, you did some stupid stuff in the 80s, or both. In any event, the solution is simple: right size your infrastructure.
All hardware gets more expensive with age. It requires more power than yesterday, it generates more heat than yesterday, and given ALL the threats, it’s less secure than yesterday.
No matter how long you have owned your hardware, it’s out of date. Run a collection. Let the data tell you what you are using. Update your hardware for what you need and give yourself some headroom for growth.
And ask yourself this question: Why in the HELL would you pay your newly anointed favorite pariah Hock Tan to license hardware that you don’t need?
Whether you renew your VMware licensing (based on right-sized infrastructure), switch to a platform that is actually hybrid-cloud friendly like Azure Stack HCI, or settle for a me too offering like Nutanix or Scale Computing, Lenovo can meet you where you want to be.
Stop blaming the 80s for the bad haircut you took well into the 90s and stop blaming Hock Tan for your over-provisioned infrastructure.
Like never before, it’s time to right size. Let’s look at the data and let’s get to work.