Despite the recent retirement announcement of Microsoft Azure Media Services from the media industry, the rise of cloud is still evident within some workflow deployments of the Media & Entertainment space (see figure 1) and is always a hotly debated topic at industry forums and events. Some broadcasters and streaming service providers have already deployed or are testing elements of their media tech stack in the cloud to enable media workflows. This can work well for some applications and use cases. However, cost, over the longer term, remains a little unknown and for some use cases, such as OTT transcoding, it can be a different challenge altogether.

At Appear, we see ourselves as a “cloud enabler” – providing ground-to-cloud and cloud-to-ground contribution solutions that enable our customers to benefit from cloud-based workflows and media delivery. However, there are specific use cases where we are part of an overall hybrid customer solution that uses both cloud/software and hardware-based applications. This “not one deployment fits all applications” approach is forcing media service providers to question the options and applicability of deployments, platforms and implementation needed for delivery and transport workflows within each part of the media ecosystem. In recent times, we’ve witnessed an interesting shift, with some of our customers transitioning from cloud-based live transcoding to on-premise equipment. This blog explores the key drivers behind this under-the-radar revolution, the challenges faced, and the benefits that come with implementing this move… against the grain.

Change Drivers

One of the primary drivers for some broadcasters and streaming providers to transition towards on-premise equipment is Total Cost of Ownership (TCO) - the overall cost of the service throughout a typical media system/solution lifespan. While cloud-based solutions can offer scalability, they are often saddled with higher operational expenditure, and this is now being realized by a number of media service operators as they trial and run their workflows in the cloud. Some cloud providers offer a range of technical and commercial models, but vs PAYG, hardware can be more cost effective. For example, some major cloud providers will build a dedicated “infrastructure” in the cloud, committed to a customer using a CapEx and OpEx model – many of the big players do this so that they match the best application to the best commercial model. However, not everyone has the scale to do this and there is the added cost to get in and out of the cloud. Instead, by investing in their own dedicated on-premise hardware, some companies are finding they can reduce their reliance on longer term expense in cloud infrastructure and gain increased control over their budgets and operations.

Another core driver for this deployment shift is security and data privacy. While many tier 1 broadcasters have large IT/InfoSec teams, for many of the smaller players this is an expense they cannot afford. This is reflected in a report from the IABM that cites “security and privacy concerns” as the second most concerning barrier to cloud adoption (see figure 2). Broadcasters and streaming media providers deal with vast amounts of sensitive, high value content and user data. With equipment deployed locally on premises, there is perceived greater control over accessibility with media related security protocols and data privacy being guided by the incorporation of IP trust boundary policies (through guidelines and recommendations outlined by SMPTE RP2129). Additionally, with the implementation of local firewalls, control of vital security is placed back in the hands of the end user. Through the management of local infrastructure, robust security measures can be implemented meaning more tangible control of adherence with industry regulations and a reduction in the risk of unauthorized access or data breaches for specific use cases and applications.

Transitional Challenges

While on-premise equipment can offer cost/operational efficiency and security advantages, scalability and content time-to-market for some use cases can become more of a challenge. With cloud-based solutions, both configuring and initiating new services with the required scaling of resources is relatively simple through instantiating instances of media processing workflows. However, on-premise infrastructure requires planning and sometimes further investment in additional hardware to accommodate increasing demand. Scalability becomes a consideration to ensure that the system can handle spikes in both content delivery and viewer demand.

Dependent upon who manages the delivery and the on-site equipment, transitioning from cloud-based workflows to on-premise equipment can require a level of in-house technical expertise. Subject to the commercial model being employed, broadcasters and streaming providers may need to invest in skilled professionals who can manage dedicated hardware infrastructure. This may involve training the existing workforce or hiring new talent with specialized knowledge, either of which can pose challenges in terms of time, resources and cost.

The Benefits

Until recently, density and space efficiency coupled with future expansion of custom hardware solutions was an issue. Now, technology and processing power has advanced to such a level that this is much less of a concern for broadcast operational centers when balancing the need for space with the benefits of running efficient media processing on premises. The power efficiency gained by hardware acceleration and the sheer density within dedicated custom solutions through technological advances are seemingly outweighing the balance of the cost for space required in operational centers.

Speaking of costs, while cloud-based solutions are commercially structured to offer variable costs based on usage, on-premise infrastructure offers long-term cost optimization over the life of the system. The upfront investment in hardware and infrastructure can therefore be offset by reduced operational expenses over time. This approach can prove more cost-effective for some broadcasters and streaming providers where more consistent and predictable workloads are present. It also allows for easier and more rock-solid budgeting when the current world micro and macroeconomics remain very uncertain.

By shifting the focus to on-premise media workflow implementations, broadcasters and streaming providers can also gain greater control and flexibility over their service and content operations. They can fine-tune the encoding, transcoding, and delivery processes themselves to meet their specific requirements and optimize resource allocation. This level of customization allows for a more tailored and efficient streaming experience for the content and services offered to their viewers as well as reflecting the brand being delivered.

Real World Examples

According to a recent report by the International Data Corporation (IDC), a global provider of market intelligence for the IT and telecommunications markets, half of all infrastructure spending comes from on-premise deployments. IDC forecasts a 2.9% annual growth rate, projecting investments to reach $77.5 billion by 2026. In the IDC survey, 70.7% of respondents with workloads in the public cloud anticipate repatriating some services to private cloud or non-cloud infrastructure within the next two years (see figure 3).

In recent months, Sappa, one of Sweden’s largest telecom operators, examined their cloud strategy and concluded that the ABR transcoding for their OTT channels was not performing as efficiently or effectively as they needed in the cloud environment. They wanted to control their own destiny and upgrade to new services such as UHD, or take advantage of better compression technology such as HEVC - all without impacting space or power and associated costs. For Sappa, the decision to migrate from cloud to an on-premise solution was largely based on the TCO of the equipment they were buying and to take back control of their operations so they could optimize their configurations and define when they needed to upgrade their services for their customers. It is important to note here that cloud is still a good fit for some ad-hoc live event instances, but in Sappa’s case, the delivery of power hungry 24/7/365 live/linear OTT channels delivered by ABR transcoding made sense at both a cost and management level.

Sappa is not the only example of how streaming media providers are migrating to on-premise ABR transcoding solutions. The same is also true for A1 Croatia, one of Croatia’s largest telecommunications companies owned by A1 Telekom Austria group. A1 Croatia also determined that the move to locally operated, controlled and managed on-premise transcode solutions for their OTT streaming services made sense in terms of their application. They found that this type of deployment would also benefit them in terms of energy efficiencies compared to other pure transcode software solutions and significantly reduce the TCO of the overall system, all the while providing them a future ready platform delivering simultaneous HEVC and MPEG4 AVC compression capabilities.

Conclusion

The importance of defining a deployment and commercial modelling strategy around the services that you want to deliver now and in the future is crucial. For some, the cost efficiencies gained by doing this could mean the difference between success and failure. The transition from cloud-based media processing to on-premise equipment for some applications and use cases, such as Sappa and A1 Croatia’s live linear transcode workflow for OTT streaming is, perhaps, marking another trend for some in the industry. By doing so, broadcasters and streaming service providers can take back control of their systems and operations, including enhanced security. They can build service configurations that reflect their own brand of content while also boosting cost and operational efficiencies. It remains to be seen how many media service providers take the step to go against the grain and change tack concerning their media processing strategies.