Blog

Cost Comparison: On-Premises Vs Cloud Computing

Written by Alain Pellen | November 25, 2020

Trying to decide between on-premises vs cloud computing for your video processing and delivery workflows? Here is how to analyze the total cost of ownership (TCO).

When deciding between an on-premises vs cloud for video delivery workflows, video streaming and broadcast companies often perform a TCO analysis to weigh the direct/indirect benefits with the price. However, adopting a cloud-based or an on-premises, or even hybrid cloud model can provide additional benefits that may not correlate with a specific dollar amount.

So, although a TCO analysis may be useful for calculating the monetary value of new IT models or hardware components, the information you obtain may be slightly misleading, especially when it comes to analyzing the costs related to delivering linear and streaming channels.

Weighing the tangibles like expenses and efficiency with intangibles like disaster recovery and elasticity or adaptability is an important step in the evaluation process for you TCO analysis. Let’s start with the basics before we review the associated advantages, potential downsides, and costs of on-premises versus cloud computing for video processing and delivery.

What is Cloud Computing for Video Processing and Delivery?

The cloud isn’t just for storage and archive functions. Many distributors are migrating most of their ecosystem (CMS, DRM, advertising, etc.) to the cloud. How though can you use the cloud for your complete video delivery chain?

Cloud computing is the practice of using a network of remote servers for on-demand access to software and data management, including storage, processing, and archiving. All the tasks you carry out or manage on a local server, personal computer, or datacenter on-site can be performed with cloud compute technology off-site.

Cloud computing service providers typically offer SaaS pricing models to customers. The service provider handles the physical aspects of running an on-premises computing location, including equipment purchases, software upgrades, system maintenance, and monitoring, and all the other tasks that managing a data center requires.

Customers, sometimes called “cloud tenants,” pay a monthly subscription based on their pricing tier and usage to “rent” the cloud service technology. For video service providers, cloud technology has opened the door to more flexible revenue models.

With the advances in cloud-native platforms for video streaming, service providers can ingest, edit, encode/transcode, stitch ads, deliver, and store video in the cloud, just as they would in a private data center.

Types of Computing Clouds for Video Processing

The two primary types of computing cloud deployment models are the public cloud and private cloud:

 

Public Cloud Computing

A public cloud is an off-premises solution for end-to-end video workflows hosted by third-party providers such as Microsoft Azure, Amazon Web Services (AWS), IBM Cloud, or Google Cloud. Customers share server bandwidth, access software, store data, encode/transcode, and more over the internet.

Advantages of the public cloud for video processing and delivery:

  • Low maintenance costs. The cloud provider handles all the expenses associated with hardware, software, infrastructure, maintenance, monitoring and more.
  • Reliability. Having a multitude of servers available in redundancy, different locations worldwide drastically minimizes the chances of failure. The advantage is that the cloud service providers can guarantee uptime for the highest availability.
  • Disaster recovery. With servers in various locations, localized natural disasters will not affect operations in the cloud. Data is secure and out of harm’s way. 
  • Friendly to remote teams. Cloud-based workflows allow global teams to run tasks remotely rather than on-premises at a single site. You can have your operational teams running your services from anywhere. These capabilities were essential during the COVID-19 shelter-in-place mandates.
  • Elasticity. Dynamically increase or decrease usage as your workload demands evolve. You only pay for the service when it’s used, rather than investing in infrastructure that you may only use when sudden traffic spikes occur. 
  • Scalability. You can request additional resources for processing power or data storage, in just a few clicks. Cloud providers have all the hardware and software resources in place to offer rapid scaling with minimum impact on your service.  Live events are excellent examples of when service providers need flexibility to scale, and with the confidence that the video and service quality remains exceptional.

 

Private Cloud Computing

A private cloud operates similarly to a public cloud. However, the hardware and software resources are only used by a single company. 

You can have a private cloud on-site running on your data center or choose a third-party service provider to run your private cloud in the cloud-provider’s data center. The first option requires you to own, manage, and maintain the hardware and software running on the data center, where the second option leverages a third-party infrastructure that is dedicated to your services.

In either case, the services and infrastructure are maintained on a private network, with hardware and software resources accessible to those only in your organization. Government agencies and financial institutions are common examples of when private clouds on private data centers are used since they demand higher control and security. 

Advantages of a private cloud for video processing and delivery:

  • Greater control. Your organization can design and customize a cloud environment to meet your specific business needs and IT requirements
  • More privacy. Since only those within your organization have access to your private cloud, you are solely responsible for your data and resources.
  • Scalability, to an extent. A private cloud on a private data center can be a challenge to scale due to costs and delivery time of the required hardware. Although a private cloud running on a private data center offers more flexibility, you're still limited by your hardware infrastructure. This means you’ve got limited resources to scale, and it’s hard to scale fast or dynamically. The public cloud has a real advantage here over private clouds running on private data centers.

 

Total Cost Comparison: On-Premises vs Cloud

A cost analysis of cloud or data centers for your video workflows includes calculations and predictions for CAPEX and OPEX expenses.

CAPEX (capital expenditures) occurs when an organization spends money to invest in new equipment, software, infrastructure, etc. On-premises data centers must take on all these expenses to get a service launched, and then they need to keep refreshing the infrastructure every three to five years. This cash drain may take away from other initiatives. 

OPEX (operating expenses) occur regularly as part of the company’s daily operations, similar to paying a utility bill. Video service providers running their channels in the cloud will be billed per usage monthly, providing a relatively fixed expense businesses can budget for in advance. 

Key questions evaluate the TCO of your video processing and delivery system:

  • How easy/expensive will it be to maintain reliability for the life of the system?
  • What will it take to uphold our desired level of service?
  • How easy and what cost is required to improve performance and enable new features and evolutions to advanced technologies?
  • When will the system outgrow the location? What would a bigger site cost to rent and operate?

What to Include in Your TCO Analysis?

Network Infrastructure and Maintenance 

On-premises data centers for video processing and delivery:

Data centers on-premises carry expenses for the material needs required to enable video processing and delivery. This includes networking gear, servers, and other hardware infrastructures.

Taking all that into account, an on-premises solution demands a budget for:

  • Renting/paying for building space and insurance
  • Utilities for electricity, cooling, UPS
  • Adding multiple server racks
  • Upgrades: adding memory and RAM to each individual server, replacing outdated CPUs with more powerful ones, software updates, etc.
  • Maintenance
  • Refresh after year five

 

Cloud Solutions for Video Processing and Delivery Workflows

With cloud you don’t need to forecast, you just scale as needed. Where for on-premises you need to forecast audiences and traffic three to five years in advance due to the logistics involved in server procurement and storage, installation and operation.

The cloud provider handles all the typical datacenter CAPEX expenses, and you won’t have to shoulder the investment or liability, which is especially useful for new and growing organizations.

 

Labor Costs

Migrating to a cloud workflow may save your company a significant amount in relation to regular labor expenses. Imagine being able to offset costs that your cloud vendors will take care of including: vetting, hiring, and employing a highly skilled IT team that you would no longer need to manage. Those teams would take care of some of these essential tasks:

  • Server storage and maintenance
  • Security patches
  • Network configurations
  • Seamless software upgrades
  • Troubleshooting

Off-loading some of these more time-consuming or labor-intensive tasks to a cloud partner prevents your in-house team from getting bogged down in maintenance. Instead, your team can work on improvements, testing new features, and other high-level tasks. Plus, you may not need to hire additional staff to handle an influx of new users with a cloud-based labor force in your corner.

 

Traffic Costs and the CDN

A content delivery network (CDN) reduces the expense and latency of transferring content to your users. It sits between viewers and your servers, making it faster to deliver cached data from multiple sources rather than constantly accessing this data from your servers alone. 

To reduce the cost of content transfers over the CDN, you can leverage the cloud. The amount of traffic to reach the CDN is the same whether the origin server is on-premises or in the cloud, but cloud vendors and large CDN have been able to colocate some of their operations so that the bandwidth to reach the CDN can be minimal or null.

 

Additional Points to Analyze Costs

The results from your TCO analysis may not reveal a clear response to which option is the most advantageous for your video streaming business. There are other factors to consider before the choice becomes apparent. Here are a few more important aspects to include in your TCO:

 

End-to-end Workflows

Playout, which involves channel origination, is only one part of the video processing and delivery chain. Other parts of the process include compression, encryption, statmux, dynamic ad insertion, and VOD creation and delivery. You can accomplish all these tasks to prep and deliver consumer-ready content in the cloud in one efficient, refined workflow.

 

Availability

On-premises data centers require robust redundancy that entails practically duplicating your physical infrastructure, which is costly to manage and maintain.

The cloud offers high availability and solutions for disaster recovery. You get the same level of performance without the excessive recovery expenses linked to typical data center economics.

 

Elasticity, Agility, Scalability, and Innovation

No one can truly predict the future of their service roadmap, and you’ll need a solution that can keep up as you adapt to evolving market demand. You can’t add infrastructure to an on-premises workflow during peak demand and return it when viewership subsides after the event. However, cloud-based workflows are elastic and the services expand and shrink with your needs. The cloud is also easier and faster to scale as your business grows.

Plus, it's often faster to launch a service in the cloud and you don’t need to spend a lot of time or capital to innovate and test new ideas. You can test and deploy new services, customized user experiences, different monetization models, and more.

Cloud workflows are also ideal for pop-up channels that run for only part of the day, week, month, or year. You only pay for the time you use the service for, making it a very economical option, compared with a hardware deployment. 

 

What about a Hybrid Cloud Computing Model?

A hybrid cloud combines the on-premises infrastructure of a private data center running a private cloud, with a public cloud service. With this model, you can securely access and move data between both environments.

A hybrid cloud is an effective way to handle spikes in viewership during live events. For example, you're delivering the video content for a championship sporting event that creates massive peaks in demand, but only temporarily. You’ll require more resources (CPU, processing power, etc.) but only for a short period of time. The public cloud is an effective and cost-efficient way to handle any overflow, and it eliminates any additional investment in on-site resources. 

Advantages of a hybrid cloud:

  • Refined data management. A common misconception is that using the public cloud requires sharing sensitive data with cloud providers. In a hybrid solution, you can keep any data you'd like in the datacenter or in a private cloud.
  • Keep existing infrastructure. If you’ve already heavily invested in on-premises infrastructure, you can integrate with cloud deployment with ease. However, you’ll need to make sure your current hardware and software integrates seamlessly, otherwise you may encounter issues that cause your workflow or apps to crash.
  • “Cloud bursting.” Organizations can rely on their on-premises infrastructure for most processing workflows, and then burst through to the cloud when demand increases. You’ll only pay for the resources used during this temporary influx rather than having to invest in the equipment and maintenance.
  • Disaster recovery and service continuity. Disaster recovery services in the cloud also enable operators to have a backup ready without paying for usage during the time that the service isn’t running on the backup channel. Pay as you use models can be considerably advantageous for video streaming providers, especially from an economic standpoint.
  • Utilizing a Hybrid Cloud model means you can also benefit from cloud-native features and a common user interface for both environments. If keeping your existing video workflow on-premises but shifting other needs to the cloud make sense for you, a Hybrid model may be beneficial.

 

On-Premises vs. Cloud Computing: What is the Right Solution for Your Business Needs?

When choosing between on-premises vs. cloud video processing and delivery, there are a wide range of technical, operational, and financial considerations to discuss with your team.

Only you can make the decision on which cloud computing model is best for your video processing and delivery workflows: cloud, on-premises or maybe even a hybrid system that includes both on-premises hardware and cloud technology.

If you’d like to discuss your specific needs with an expert, click here to learn more about what the Harmonic team can do for you.

Seamless software upgrades

Troubleshooting
Off-loading some of these more time-consuming or labor-intensive tasks to a cloud partner prevents your in-house team from getting bogged down in maintenance. Instead, your team can work on improvements, testing new features, and other high-level tasks. Plus, you may not need to hire additional staff to handle an influx of new users with a cloud-based labor force in your corner.
A public cloud is an off-premises solution for end-to-end video workflows hosted by third-party providers such as Microsoft Azure, Amazon Web Services (AWS), IBM Cloud, or Google Cloud. Customers share server bandwidth, access software, store data, encode/transcode, and more over the internet.