Ask an Ericsson or Huawei executive about open RAN technology and the habitual response is that it does not stack up against their traditional systems on performance or cost. But a breakthrough that should have been demonstrated by Deutsche Telekom and Intel at this week's aborted Mobile World Congress will go a long way toward silencing that argument.

Right now, the criticism is valid, says Alex Choi, Deutsche Telekom's senior vice president of strategy and technology innovation. One of the open RAN goals is to replace the customized gear used in mobile networks with general-purpose equipment. This would support the new software being developed by open RAN companies. But the standardized hardware remains a poor substitute, says Choi.

"The reason Ericsson, Huawei and Nokia are using more customized chipsets is that they give you a cheaper solution," he tells Light Reading. "If you have an ASIC [application-specific integrated circuit] for baseband processing, it is always cheaper than using a general-purpose processor [GPP] like an Intel processor."

A potential answer is to offset the GPP performance costs in other areas, and that is exactly what Deutsche Telekom and its partners have been doing. The German telecom giant teamed up with Intel as well as VMware, a software company that has specialized in developing products for Intel's architecture. The system they built, which specifically runs on Intel's FlexRAN architecture, allowed them to fully separate the radio software management from the underlying hardware.

VMware's platform was critical in all this, according to a Deutsche Telekom spokesperson, because it was fully compliant with new open interfaces developed by the O-RAN Alliance, an industry association. That meant Deutsche Telekom could turn the radio resource management into a discrete application sitting on top of this VMware platform. The next step was to look at making improvements in the management area.

"The cost of using the GPP is more expensive than an ASIC, but that can be offset by the higher layer – this advanced radio resource management algorithm," says Choi. "In total the overall cost can be lower than in the case of using a traditional platform."

The algorithm Deutsche Telekom is using in its demonstration was developed by Cohere Technologies, a Californian software vendor whose CEO Ray Dolan recently spoke with Light Reading in some detail about his company's products. The important thing here is that Cohere's algorithm led to a major improvement in radio performance during Deutsche Telekom's trials. Indeed, Choi says throughput doubled on the operator's network, leading to a reduction in total costs.

This endorsement by one of the world's biggest telecom operators is obviously fantastic news for Cohere, but Choi expects rivals to emerge with similarly innovative products. "Once we open up this platform, there will be others," he says. "That is a key benefit of this kind of open platform. Now you are inviting all the innovators."

Want to know more about 5G? Check out our dedicated 5G content channel here on Light Reading.

News of the breakthrough comes days after one of Ericsson's top executives took aim at the use of GPPs in radio networks. "We spend significant amounts of R&D to deal with the compute power needed to handle massive MIMO [an advanced 5G technology] and currently general processing cannot handle those types of workload," said Fredrik Jejdling, the Swedish vendor's head of networks. "So far we have not seen any technology that can deal with performance efficiency to the extent that ASICs can."

Choi is sympathetic, and the technology developed with Intel, VMware and Cohere – which is undergoing testing and validation at Deutsche Telekom's headquarters in Bonn, Germany – does not mean the industry has stopped looking at ways to improve GPPs. "The performance, reliability and cost structure are not quite there yet," he says. "The technology is progressing well but still not meeting all these key requirements from the operator side."

Nevertheless, if Deutsche Telekom can build support for the technologies it is now testing in Bonn, some of those arguments about the shortcomings of open RAN may start to lose their weight.

Related posts:

— Iain Morris, International Editor, Light Reading