A Hybrid Approach Guides the Changing Face of On-Prem Encoding
Ten years ago, a live or video-on-demand (VOD) encoder ran on a standalone computer located in a server room or in a corner of your studio, connected to other encoders or controllers via a local area network. Today, as computing resources are consolidated into clusters of virtual machines, or private or public clouds, encoding customers are seeking new deployment options to match their changing computing architectures, and encoding companies are responding.
To understand the specifics of this trend, I reached out to about 20 encoding companies worldwide and spoke to twelve vendors (some off the record) that serve a broad swath of live and VOD encoding markets. On one level, I learned that in many markets, both large and small, the traditional “encoder as a standalone appliance” model still dominated. On the other, I learned that all vendors are offering new deployment and licensing models, some in response to customer demands, some on the “build it and they will come” plan.
These discussions led to the following conclusions.
Markets Are Changing, But at Different Speeds
Not all companies are virtualizing their computing infrastructures, of course, and not all that are virtualizing are transitioning away from traditional encoding schemas at the same pace. For example, Elecard sells multiple encoding products, including live encoder CodecWorks (which was used to stream the FIFA Confederations Cup in 2017) and VOD encoder Converter Studio Pro.
According to product manager Vadim Blinov, most CodecWorks users install the software in a single on-prem computer because they care more about performance, while scalability is achieved with more blades for high-density hardware configurations. In contrast, with Converter Studio Pro, Elecard is seeing demand for a cloud version, which is now under development. Explaining the difference, Blinov comments, “Many customers want access to the cloud to manage spikes to their VOD demand. We’re not seeing the same level of demand volatility in live encoding.”
Similarly, Comprimato’s Live transcoder runs on any standard x64 server device with GPUs, or it can be virtualized in a data center via Docker (Figure 1). Although the company is experimenting with a cloud version, it hasn’t deployed that. According to CEO Jirí Matela, “Customers we talk to aren’t ready to abandon the old-fashioned model of controlling equipment on-prem. Though they’re asking about the ability to run in a cloud or use as a service, they don’t appear ready to make that transition.”
Comprimato’s Live transcoder can be installed in a server or in a virtual machine.
Anevia’s Jérôme Blanc described how the market was changing for his company’s Genova Live encoder. “Our traditional markets are broadcasters and telcos. Broadcasters typically have a limited number of live TV services, are comfortable with appliances and a CapEx model, and haven’t seen the benefit of virtualization. Telcos offer more channels, were OK with appliances, but 2 to 3 years ago wanted software-only pricing. Some have tried virtualization, but discovered a negative performance impact, probably because it was based on virtual machines at that time. This group is now starting to ask about OpEx rather than CapEx models.
“Since early 2018, we see a new type of customer: OTT platforms, who want to aggregate and offer live encoding and packaging services to broadcasters, and telcos,” Blanc says. “They build their infrastructure from the ground up on virtualization, using containers for maximum performance instead of VMs, and either host it themselves (on-prem or rented space in an external data center that they manage) or trust a public cloud to host it (AWS, Azure). They need elasticity, because they can’t predict how many live TV services they will have to handle.” According to Blanc, this class of customer is also pushing for an OpEx model because that’s what it’s offering to its clients.
In the VOD space, I spoke with Ikuyo Yamada, CEO at Capella Systems, which sells a VOD program called Cambria FTC that runs on Windows. As described by Yamada, although most of Capella’s clients run Cambria in a traditional on-prem workstation, about 30 percent are now installing in virtual machines or in the cloud, which Capella supports with a floating licensing model that lets customers activate and deactivate Cambria installations to meet their worldwide needs. The next update will also add load balancing to the cloud.
According to Yamada, Capella hasn’t seen a lot of pressure to change from a CapEx to a usage-based model. “Our customers buy their encoders for a relatively fixed demand of VOD processing,” she says. “In these cases, a traditional licensing model is less expensive than most usage-based models.”
At the other end of the spectrum are companies like Harmonic, which has expanded its product line to support most possible deployment options. Although products like the Electra 9200 are still sold as appliances, or as software with custom off-the-shelf (COTS) hardware (as Electra X2 or X2S), it’s also available for deployment in virtual machines (Electra XVM) or containers (Electra Docker), and even via SaaS with usage-based pricing (VOS 360) or a full software stack installable in a customer’s private or public cloud (VOS SW Cluster). The hardware and software products are sold under traditional licensing models, while the SaaS enables companies to access the same technology with per-use pricing. The same encoding technology is used in all products.
The Cloud Is Coming On-Prem ...
As corporate encoding environments become more cloud-like, traditional cloud-encoding vendors like Encoding.com and Bitmovin are offering their solutions for deployments inside the firewall. Gregg Heil, CEO of Encoding.com, explains why several of his largest clients opted for this method. “As you can imagine, some of our largest media customers have massive compute and storage investments that they want to use,” he says. “However, they also want the ability to elastically scale as needed, access critical video workflows, and to avoid the headaches and costs of managing traditional in-house software and hardware encoders.
“We can run the Encoding.com service in any data center. When running our Hybrid solution in customer data centers, we update all the engines to ensure the solution is always current,” he says. “Unlike old-school software encoders, when the processing is finished, customers can seamlessly repurpose the CPU resources for other functions.” As described by Heil, the Encoding.com individual applications exist in containers, driven by an orchestration system that can keep all content inside the firewall for security-conscious customers. Other clients, who don’t see encoding in the cloud as a security concern, can burst to the Encoding.com public cloud when on-prem compute resources are 100 percent used to complete all encoding jobs as quickly as possible.
Similarly, Bitmovin provides its encoding software on-prem behind its customers’ firewalls, with overflow capabilities to public cloud environments, delivering many of the same benefits as Encoding.com (Figure 2). According to Bitmovin vice president of solutions Igor Oreper, since the Bitmovin encoding service can run on customers’ own infrastructure, software pricing is a fraction of that charged for the SaaS-based service. Oreper also advised that the public/private cloud native nature of the Bitmovin content processing service allows its customers to encode where it is most efficient for final delivery and most cost effective when considering total cost of ownership, including network egress.
Bitmovin offers its encoding services installation behind the firewall and as a public SaaS.
... and On-Prem Is Moving to the Cloud
Most original cloud encoding services were just that: encoding services that handled the last stage of media creation and deployment, plus encoding and packaging, but none of the antecedent workflows and none of the subsequent distribution roles. As cloud services push their encoding capabilities into the enterprise, traditional encoding vendors are porting the full functionality of their in-house workflow systems into the cloud. This allows customers to push more of their in-house workflows into the cloud.
Streaming Media West provided a look at new options for demanding live cloud transcoding jobs. Here's a rundown for those who couldn't be there.
We wrap up our survey of the encding and transcoding landscape with looks at Verizon Digital Media Services, Media Excel, Comprimato, Elecard, Capella Systems, Epic Labs, EuclidIQ, and NGCodec
We continue our survey of the encoding & transcoding market with looks at Encoding.com, Bitmovin, Brightcove, Beamr, Cisco/Synamedia, and Ericsson/MediaKind
Encoding and transcoding are at the heart of every OTT and online video workflow. The first article in this three-part series gives an overview of the technologies and a look at three major players in the space: Harmonic, AWS Elemental, and Telestream.
Now widely available, per-title encoding makes whatever codec publishers are already using more efficient by creating a custom optimized encoding ladder.
An insightful new service called Mux Data makes quality of experience monitoring and analysis easy. This illustrated guide explains how to use it when diagnosing problems big and small.
If you're not already using per-title encoding, it's time. Here's a guide to choosing the tool that's best for you.
Companies and Suppliers Mentioned