Skip to content
Equnix Metal

Access to Server Hardware is Disappearing. Does it Matter?

Written by:

 The act of breaking stuff and having to fix it, of taking it apart and putting it back together again, takes confidence you can often only get through a lifetime of tinkering.

Last month I was lucky enough to welcome a good friend, Sam Machiz, to the ranks of Packet. Like me, Sam has spent the majority of his adult life in the web hosting industry (he started Penguin Hosting while still in high school took me until my senior year of college to get into the weird, wild world of servers, datacenters, and networks). We’re both tinkerers, gadget nerds, and bleeding-edge early adopters who can’t help but ask dozens of follow-up questions when we hear about a new widget.

Since enticing Sam into Packet, he and I have spent countless hours geeking out on the hardware questions facing our world: new server designs, open networking concepts, 5G, programmable FPGA’s, vector processing, reverse RAM-drives (drive RAMs?), JAVA heap sizes, network interconnection options, and various processor architectures. Aside from helping me tackle the day to day of choosing our next server platform, it’s been just plain old fun - rekindling the native sense of curiosity that got me into this business.

The Natural Philosopher in All of Us

Many of us shy away from things we don’t know very well - or aren’t specifically trained to be an expert in. This is an especially easy trap with infrastructure, with its ever-changing and highly complex layers. You learn pretty quickly that the only thing you know for sure is that you don’t know it all.

What I dig about Sam’s approach to infrastructure is that he’s not scared of any topic. His curiosity about what’s under the hood takes precedence over any need to appear “smart”. It's an approach to technology that reminds me of those Royal Society chaps I first read about in Neal Stephenson’s Baroque Cycle. Isaac Newton and his buddies weren’t interested as much in impressing their dinner party friends as they were about trying to understand everything they could about the (magical, mysterious, complex) world around them.

The MacGyver Effect

But how do you get to be unafraid of technology and all its physicality and intangibles?   I think it has to do with hacking, or better yet: tinkering. The act of breaking stuff and having to fix it, of taking it apart and putting it back together again, of pulling out some McGyver moves when things don’t have an instruction manual.

MacGyver photo.

This is how the best innovators out there have come to be. For instance, great coders learn to create amazing solutions by breaking some code and fixing it and making countless mistakes along the way, finding in the process their own opinions and ideas.

But with physical technology (such as servers, networks and datacenters) most people have very little access to the raw gear with which to tinker.  The primitives that helped me and Sam - and most veterans of the cloud infrastructure industry - are disappearing from daily life, replaced by MacBook pros and elegant APIs to 'virtual' clouds of compute.

For Me, It Started With Building PC’s

Photo of mainboard with heatsink fan.This past week at the Packet office, we hacked on a test board for a new processor we’re evaluating. This is preproduction, doesn’t-have-a-BIOS kind of stuff.

It required several adapters to get serial output to a fancy Macbook Pro (thanks Adam Rothschild for the WiFi console adapter!), meddling with dip switches, a paperclip, and plenty of phone calls back to the engineers who built the thing. In the end, we got it booted with Ubuntu and have been actively abusing it for the past few days.  We mucked around so that we could analyze the relative performance of a brand spanking new ARM chip against our current Intel Atom chips.

The first time I ever did this was when my aunt Kathleen (a proud investor in Packet!) brought home a brand new IBM PC Jr from her job at IBM. At some point when nobody was looking, I opened it up and poked around. It didn’t have a hard drive, so I for sure wasn’t going to be upgrading that! However, I did get a feel for things like PCB boards and a few years later graduated to building my own PC’s, crimping my cables and figuring out how to get all the wires plugged into the right parts on an ATX motherboard.

Funny enough, this prepared me handsomely for the hosting business, where you are constantly context switching: from playing with random new hardware boards (see above) to reading for memory compatibility details, to testing out the latest Ansible release. My youthful experience of messing with the physical technology removed a lot of mystery and gave me the confidence and passion for figuring this stuff out.

In fact, it’s not too different from my father’s generation - just different tech.  They spent much of their youth in the garage, changing the oil or a timing belt or a tire on a 1967 five-lug Mustang or VW bus. I was always amazed that my dad was so comfortable under the hood of a car. Heck, he grew up futzing with them!

By the time I was ready to drive, you couldn’t even access the engine of most cars without special tools, as it was all hidden somewhere underneath a beautiful molded cover.  Now in my 30's, most of us just lease a car and get new ones before we needed to figure these things out. No wonder I would rather mess with a motherboard than a transmission!

Sounds like a modern MacBook to any of you?

The primary computing device for most younger people is a Mac or mobile phone, where you physically can’t take it apart (or wouldn’t dare to do, as it would void your warranty). Combine that with the disappearance of neighborhood ISPs (who back in the day let you rack a few things and muck around) and I argue that innovators in the hardware stack are going to be much harder to come by in the years to come.   

Of course, there will be exceptions and there will be experts that will push hardware forward at the most fundamental levels. Raspberry Pi fanatics push those little computers to amazing heights, for instance. But the accidental hacker? The tinkerer turned entrepreneur? I think that is where we’ll see the gap.

So...does it matter?

Innovating the Future means Hacking the Full Stack

I think the “full-stack” is more than just backend and frontend web stuff.

Some of the most exciting technologies that will affect our lives and planet face seemingly insurmountable challenges if accomplished via code alone. Just think of buzz words like IoT, self-driving cars or virtual reality!   Will the massive technology, financial and environmental challenges that surround these "things" be solved all in code sitting above the Linux kernel? Or will new architectures, offload cards, form factors, networks, and datacenter environments be part of that too?

From what I see each day, I’d say the answer is most definitely 'yes.'  But it won’t happen because some fancy new FPGA offload card shows up in the market. It will be because a large enough group of curious and status-quo challenging developers, entrepreneurs, researchers, hobbyists, hackers, tinkerers, and consumers get to innovate with them.

In a world where the cloud has come to mean a series of abstracted API’s that move you so far away from the hardware you can barely recognize it exists, I’m proud to be in the business of providing access to raw hardware to the modern hackers of the world. At Packet, we get up each day to make it easier for developers to leverage powerful infrastructure while being able to “touch” the raw bits.

My hope is that we’ll be the hardware petri dish for a new generation of “Sam’s”. I can’t wait to see what they all come up with.

Like what you see?

Then let’s get you started on your journey with Equinix Metal