Tantra Analyst

HPE and Qualcomm announcement signifies the role of cloud and telco collaboration in Open RAN

The talk of the outsized role of cloud players in Open RAN is as old as talk of Open RAN itself. After prolonged anticipation, the collaboration between traditional telco and cloud players is finally coming to fruition. The recent HPE and Qualcomm announcement about virtualized Distributed Unit (vDU) is an important milestone and sets off an industry trend. It also signifies how both industries have to come together for Open RAN to succeed. 

Collaboration between cloud and telco players is critical for Open RAN 

A fundamental tenet of Open RAN is cloudification. Not only employing cloud architecture but also utilizing cloud infrastructure to simplify and optimize the cost of RAN deployments. That means using cloud-native architectures such as Kubernetes, Hypervisor, and Microservices for Open RAN. These make RAN completely software-defined and just another workload that runs on the cloud infrastructure. Cloudification brings many benefits, including RAN to rapidly and cost-effectively scale up, following the trajectory of the cloud market. As well as expand the telco ecosystem to giant cloud players with deep knowledge and significant financial muscle.

However, there is a slight kink in all of this. RAN workload is unlike any other cloud workload. It is incredibly complex and highly latency-sensitive. In fact, that is the reason why it has taken this long for the Open RAN to be a commercial reality. The last few years of work have convinced everybody in the know of two things. One, it is impossible to run the full RAN workload efficiently only on generic compute (aka COTS – Commercial Off the Shelf), requiring specialized accelerators. Second, the need for extensive domain expertise that only key telco players can bring. In such a case, the collaboration between cloud and telcos players is inevitable.

Leveraging best of the breed 

The success of Open RAN depends on matching or exceeding the performance of legacy systems. Early adopters indicate that the performance of relatively simple configurations such as 4T4R and Rel. 15 based enhanced mobile broadband feature set is at par or better than legacy networks. So, now the race is to prove that the networks with advanced antenna configurations, wider bandwidths, and evolutions beyond Rel. 15 can achieve that while still delivering on the cost-savings promise. The industry will need best-of-the-breed solutions from both cloud and telco systems to accomplish that. That is precisely what this latest collaboration between HPE and Qualcomm achieves.

Ever since announcing its 5G RAN solutions back in 2020, Qualcomm has been making waves, getting endorsements from a long list of leading and large operators, including Vodafone, Dish, Reliance Jio, Airtel, and others. 

Qualcomm’s X100 5G RAN accelerator is a critical building block for any Open RAN system. It offloads latency-sensitive, real-time functions such as demodulation, beamforming, channel coding, Forward Error Correction (FEC), and others, often referred to as “Layer-1 functions” from the host processor. It works in “Inline” mode vs. the other option, “Look aside” mode. Inline mode is a must to achieve the extremely latency-sensitive workload, especially for configurations like 64T64R, bandwidths of 100MHz or more, features such as Carrier Aggregation, and applications that need ultra-low latency, and ultra-high reliability (aka URLLC – Ultra Reliable Ultra-Low Latency Communications).

Please check out the article A fresh look at building 5G radio access networks for more details on Inline accelerators.

Qualcomm, a leader and a dominant 5G technology player brings decades of radio experience with X100. HPE, on the other hand, has been a proven leader in the cloud market. Their HPE ProLiant DL110 Gen10 Plus is a carried-grade server platform optimized for RAN workloads. 

The X100 card can address various traffic 5G deployments, either in a single site or aggregated vDU configurations. It supports the popular PCI-e interface and can work with any cloud/compute platform that supports that interface, including x86 Xeon processors. 

The announcement claims this vDU can provide up to 60% lower Total Cost of Ownership (TCO) improvement over today’s vDU solutions, based on both vendors’ joint study. If true, that indeed is an impressive achievement. I can’t wait to get my hands on that study to examine the details. 

Beginning of an industry trend 

So far, Qualcomm and other telco players have been working with others within the industry. But this collaboration is going beyond the traditional partners and reaching a major cloud player. That indeed is a significant development and will be a recurring theme as the march toward Open RAN progresses. With reinvigorated Mobile World Congress upon us, after two years of subdued presence, I am sure there will be many more such announcements in the lead up to and at the event, following this trend.

Meanwhile, to read articles like this and get an up-to-date analysis of the latest mobile and tech industry news, sign-up for our monthly newsletter at TantraAnalyst.com/Newsletter, or listen to our Tantra’s Mantra podcast.