Building out your own on-premise infrastructure offers control, while cloud-based infrastructure offers efficiency and convenience, so the choice comes down to a strategic decision based on your company's needs
While some organizations have the budgets, technical acumen and strategic priorities that mandate on-prem solutions for video delivery, very few organizations actually choose to build their own video delivery infrastructure. Why? Video is resource intensive and most organizations aren’t delivering dozens, or even hundreds, of terabytes of video a day and therefore can’t drive the economies of scale to make it a worthwhile investment.
The reality is that it takes a lot of money, effort, and time to build out and maintain a network with the appropriate capacity to deliver high quality video around the globe. A key advantage of operating media delivery as a single service is that it becomes easier to spread costs across a large number of customers and make more efficient use of expenditures than any one of those customers could do on their own.
Most organizations have audiences that are large enough that a private infrastructure would create a fairly high degree of complexity in a video deployment model. For smaller organizations with a few hundred viewers who are geographically concentrated, the economies of scale are often more achievable and make good business sense, but those same organizations often lack the technical resources needed to build out their own video delivery infrastructure and frequently opt for freely available services like YouTube and Facebook Live.
Even an audience that’s small but geographically dispersed adds complexity. You’d only need a few points of presence if video quality is not a priority for your audience; that changes dramatically when you need to deliver high quality video and would require several points of presence in various locations based on where your audience is most likely to be located.
As your audience grows from a few hundred to low thousands, you start to reach a point where you need multiple compute resources combined together in a load balanced and pooled manner to service that load. This also introduces the complexity of managing the orchestration of that infrastructure. How much capacity do you need? How many machines? In which locations? Which machine will get the next request? What if one goes down?
If you want to build a video delivery system that's capable of supporting a large audience, you need to consider the setup, configuration and management of that infrastructure. And, regardless of how robust the tools are or how preconfigured the solution is, there’s still a significant amount of time, effort and energy, and people needed to set it up, to monitor it, to operate it, and make sure it’s available for all of your events.
Some organizations may choose to build out their own private infrastructure because their security policies dictate a “Walled Garden” network so considerations like economics and overall workforce efficiency are set aside. We’re starting to see that change, as even the US government is beginning to use cloud-based services.
With its private fiber network and industry-leading encryption, both in transit and at rest, well-designed platforms like Phenix are able to address many of the security concerns that underlie the decision of a private build out.
Another important consideration is the need for a high-availability, fault-tolerant system.
Attempting a private build-out for high-stakes applications where crashes and stuttering are unacceptable would be tremendously difficult and expensive.
For example, industry best practice is to avoid any one single point of failure, which is why Phenix employs a multi-cloud approach, partnering with big industry players like Google and Oracle - and we’ve architected our platform such that the infrastructures for these hyperscalers are providing redundancy for one another.
To achieve the level of fault tolerance provided by Phenix in a privately-deployed, self-managed system, you would need to make sure you’re sufficiently staffed to negotiate contracts with multiple cloud vendors that offer five 9s service level agreements, manage deployments across cloud vendors, make sure that those pieces of software can communicate with one another about the availability of the other, and route traffic from one to the other in the case of failure. If you want high-availability with no single point of failure, you're going to have to build a complex system architecture, operate it, manage it, test it, and make sure that it’s working properly.
The good news is that Phenix has already done that - and we've done that on top of big players like Google and Oracle. And, Google and Oracle already make huge efforts to make sure that their platforms don't go down. Additionally we’re making sure that even if they do go down, which is highly unlikely, we remain online. So, by definition, we are more fault-tolerant than Google or Oracle because we use them both.
When working with a platform partner, you benefit from the industry’s best services including compute services, database services, storage services, delivery services, and the list goes on – without having to negotiate those contracts or configure, manage, and maintain those services. You don’t have to replicate Operational Excellence procedures when it comes to hardware, maintenance, failures of hard drives, failures of power supplies, etc.
For lower stakes events at organizations without the technical expertise to build out video delivery infrastructure, there are some great free options like YouTube TV and Facebook Live.
Using a platform as a service allows you to focus on innovation and building a product that provides value to your customers. It's always-on, fault-tolerant design reduces your go-to-market risk.
You just turn it on when you need it. You don’t have to design an entire video delivery ecosystem, maintain it, and figure out how to scale it yourself. You’re trying to bring innovation to marketing – that’s hard enough. Why burden yourself with software updates and racking servers and rolling out new versions? Let us take care of that, and you can use it when you need it.
What could be simpler than that?