Building a Cloud Infrastructure from the Ground Up

Jon Walkenhorst

5/13/21

A wise Roman philosopher by the name of Seneca once said, “Luck is what happens when preparation meets opportunity.”

Back in 2011, I was transitioning from the video games world to Hollywood and into video streaming. The new role was to deliver a cloud based transactional video-on-demand (TVOD) service to compete with much larger established players, only to see the market adopt subscription video-on-demand (SVOD). In spite of this market shift, we were successful in building a cloud based service at a time when infrastructure was still built around and deployed on purchased iron and commercial off-the-shelf (COTS) servers. The silver lining in the unexpected consumer shift towards subscription based content was the organic knowledge acquired to build a cloud native based marketplace. As a direct result, my team and I were asked to take this experience and pivot to taking on the full enterprise cloud strategy to retool and refactor three unrelated different business units. The business strategy was to move from a CapEx into OpEx infrastructure model, and what we didn't know at the time (foreshadowing here), a retool and update to business models and workflows. We made great progress in areas of machine learning, predictive analytics, digital asset management, and even cloud VFX render, only to bump against a glass wall of change resistance. What I learned is that the cloud enables technology to adapt and evolve in a way that was not previously possible, but the cloud alone doesn’t overcome the barriers to adoption.

In 2013, the CapEx to OpEx shift presented a number of barriers to adoption. Each was harder than the last to surmount. The first, data storage, being the easiest in this context moving from appliance based storage devices to object storage on COTS hardware to store assets. To content owners, this step is innocuous. The next hurdle, a little more challenging, comes from connecting physical users to cloud based virtual machines colocated with the data in the first step. This step works, but often doesn’t scale in a global way when data resides inside protected company networks and users can be anywhere. This brings us to step three, moving off premise to a hosted, unknown infrastructure. At the time, it was a terrifying concept to studios and asset creators to let their IP (original video content) be uploaded to a public cloud company. Public cloud was one of the few solutions that could keep up with the totality of content data being created daily. Data is ephemeral, compute utilization ebbs and flows, especially in Hollywood where the workload is based on seasonal factors, large versus small clients, and long-form versus episodic. These same factors and requirements make content creation an ideal use case for public cloud and the OpEx model. In the old model, while private cloud can be elastic and familiar to IT departments, it is still paid for by CapEx budgets.

Traditionally, Hollywood and creative folks have a special presence in the mind that allows them to remain focused on creation, but not necessarily welcoming of changes to their workflows. As the industry held onto their infrastructure practices in 2013, the creative mindset would have their day as well. This change resistance was pervasive from the set to the bay for many years, and when new options for saving time, money, reduced travel, and quality would come knocking, people would listen and then go back to what worked. Just a few quarters ago, the physical, hands-on, creative process workflows were both ambient and ubiquitous, and thought immutable. 

But every once in a while, a force majeure will stride through an open door, removing the impediments to change and Covid did just that: locked away the old practices of creative production. Without access to physical facilities and resources, creative professionals were required to relearn how to do their jobs from a process that hadn’t changed in 30 years or more. It was this resistance that was compelled to fall when global changes came through the pandemic door. Potentially fast-tracking adoption by as much as 5 years of organic mind shift according to industry watchers like Movie Labs.

Starting as far back as 2014, a notable number of new tools, platforms, and services (built on a hybrid of physical and virtual technologies) came on the market as alternatives to tried and tested physical processes in production. However not every hopeful platform and idea believed blending old and new was the right way to go. Not every idea is a disruption and not every impediment is technical: incumbent tools and processes are a difficult social engineering challenge. These same tools, in 2020 suffered a potentially fatal blow that would only be recovered through the availability of virtual collaboration solutions, and those I am aware of are built in a cloud native approach.

Even as recently as 2017, this idea of a cloud native solution was little more than objects and words on whiteboards. The foundational technology had been in-use in other markets, but for M/E and creative collaboration use cases, the nascent growth required visionary leaders to willingly take a few risks and have the perseverance to wait for luck. I don’t believe luck is a coincidence, but I do believe building a collaborative SaaS service for an industry and market (that doesn’t know it can benefit from a cloud native solution) is a matter of timing.

Jon Walkenhorst

Website
Jon Walkenhorst is the Chief Technology Officer at Evercast, where he works to build solutions that anticipate market shifts and identify business opportunities that help build enterprise value. Jon is a master craftsman, builder, educator, integrator, and advisory delivery partner. His experience spans product innovation, CPE, video delivery, cloud transformation, global IT, and Telco NFV/SDN.

The latest from Behind the Scenes at Evercast, delivered straight to your inbox.

Success! You have signed up for our latest articles.
Oops! Something went wrong while submitting the form.

We’ll never send you spam - we promise.

Back to Guides

Building a Cloud Infrastructure from the Ground Up

Jon Walkenhorst

5 min read time

A wise Roman philosopher by the name of Seneca once said, “Luck is what happens when preparation meets opportunity.”

Back in 2011, I was transitioning from the video games world to Hollywood and into video streaming. The new role was to deliver a cloud based transactional video-on-demand (TVOD) service to compete with much larger established players, only to see the market adopt subscription video-on-demand (SVOD). In spite of this market shift, we were successful in building a cloud based service at a time when infrastructure was still built around and deployed on purchased iron and commercial off-the-shelf (COTS) servers. The silver lining in the unexpected consumer shift towards subscription based content was the organic knowledge acquired to build a cloud native based marketplace. As a direct result, my team and I were asked to take this experience and pivot to taking on the full enterprise cloud strategy to retool and refactor three unrelated different business units. The business strategy was to move from a CapEx into OpEx infrastructure model, and what we didn't know at the time (foreshadowing here), a retool and update to business models and workflows. We made great progress in areas of machine learning, predictive analytics, digital asset management, and even cloud VFX render, only to bump against a glass wall of change resistance. What I learned is that the cloud enables technology to adapt and evolve in a way that was not previously possible, but the cloud alone doesn’t overcome the barriers to adoption.

In 2013, the CapEx to OpEx shift presented a number of barriers to adoption. Each was harder than the last to surmount. The first, data storage, being the easiest in this context moving from appliance based storage devices to object storage on COTS hardware to store assets. To content owners, this step is innocuous. The next hurdle, a little more challenging, comes from connecting physical users to cloud based virtual machines colocated with the data in the first step. This step works, but often doesn’t scale in a global way when data resides inside protected company networks and users can be anywhere. This brings us to step three, moving off premise to a hosted, unknown infrastructure. At the time, it was a terrifying concept to studios and asset creators to let their IP (original video content) be uploaded to a public cloud company. Public cloud was one of the few solutions that could keep up with the totality of content data being created daily. Data is ephemeral, compute utilization ebbs and flows, especially in Hollywood where the workload is based on seasonal factors, large versus small clients, and long-form versus episodic. These same factors and requirements make content creation an ideal use case for public cloud and the OpEx model. In the old model, while private cloud can be elastic and familiar to IT departments, it is still paid for by CapEx budgets.

Traditionally, Hollywood and creative folks have a special presence in the mind that allows them to remain focused on creation, but not necessarily welcoming of changes to their workflows. As the industry held onto their infrastructure practices in 2013, the creative mindset would have their day as well. This change resistance was pervasive from the set to the bay for many years, and when new options for saving time, money, reduced travel, and quality would come knocking, people would listen and then go back to what worked. Just a few quarters ago, the physical, hands-on, creative process workflows were both ambient and ubiquitous, and thought immutable. 

But every once in a while, a force majeure will stride through an open door, removing the impediments to change and Covid did just that: locked away the old practices of creative production. Without access to physical facilities and resources, creative professionals were required to relearn how to do their jobs from a process that hadn’t changed in 30 years or more. It was this resistance that was compelled to fall when global changes came through the pandemic door. Potentially fast-tracking adoption by as much as 5 years of organic mind shift according to industry watchers like Movie Labs.

Starting as far back as 2014, a notable number of new tools, platforms, and services (built on a hybrid of physical and virtual technologies) came on the market as alternatives to tried and tested physical processes in production. However not every hopeful platform and idea believed blending old and new was the right way to go. Not every idea is a disruption and not every impediment is technical: incumbent tools and processes are a difficult social engineering challenge. These same tools, in 2020 suffered a potentially fatal blow that would only be recovered through the availability of virtual collaboration solutions, and those I am aware of are built in a cloud native approach.

Even as recently as 2017, this idea of a cloud native solution was little more than objects and words on whiteboards. The foundational technology had been in-use in other markets, but for M/E and creative collaboration use cases, the nascent growth required visionary leaders to willingly take a few risks and have the perseverance to wait for luck. I don’t believe luck is a coincidence, but I do believe building a collaborative SaaS service for an industry and market (that doesn’t know it can benefit from a cloud native solution) is a matter of timing.

Subscribe to get our best content in your inbox

Success! You have signed up for our latest articles.
Oops! Something went wrong while submitting the form.