MWC Shanghai 2023 was dominated by promising enterprise applications and reports of the region’s significant progress towards 5G Advanced, including private networks and connected vehicles, striving to foster a unified technology for an empowered 5G global future.
The big question going into MWC Las Vegas was which themes would prevail in a market still seeking the highest value opportunities in 5G from the North American perspective.
The buzziest booths were big carriers pitching business offerings, including private wireless network plays. AI hype was at peak level—as I walked the floor, I heard someone express to a colleague how pleased they were that the term “generative AI” had been incorporated into a presentation they’d just given.
Let’s get right to some of the key takeaways we picked up across booth marketing, panel sessions, and informal discussions.
Private 5G wireless network opportunities are expanding
Security and quality of service (QoS) are proving to be key drivers for enterprise adoption and implementation of private 5G mobile networks, according to a Spirent survey of enterprises released ahead of the show. Those were both topics of interest in Las Vegas.
The combination of private 5G, edge, security, QoS/network slices, and devices has created compelling end-to-end offerings that service providers are excitedly presenting to enterprises.
In a fireside chat, Verizon highlighted its Enterprise Intelligence portfolio which includes a private 5G network, network as a service, and 5G edge. Kyle Malady, EVP & CEO of Verizon Business, stressed the need to think of private networks as “connectivity and compute.” Private multi-access edge computing (MEC) will be bundled with many private networks to provide data sovereignty, security, and low latency. That becomes particularly important in an AI future where data will be a key enabler for many enterprise use cases.
T-Mobile focused on enterprise security requirements aimed at helping businesses and government organizations strengthen defenses against cyber threats. T-Mobile’s SIM-based secure access service edge (SASE) provides always-on device authentication to increase connection security and provides SASE customers with incremental security options via network security slice capabilities.
Another network slice solution was hinted at by SVP AT&T Connected Solutions, Hardmon Williams, who discussed a new QoS on-demand capability as part of their advanced network services which would offer service level tiers for specific use cases.
There were also developments on the private network device front. To facilitate private 5G adoption, NTT and Qualcomm are partnering to accelerate the development of 5G devices and the ecosystem. Together, the companies will provide devices for private network use cases, including AR headsets, computer vision cameras, and sensors. This is encouraging news for private 5G wireless’ monetization prospects, as one of the main factors delaying private network adoption has been a lack of relevant devices for enterprises.
Of course, the devil is in the details. Will these new capabilities deliver the security, reliability, and performance QoS required by early use cases? Recent Spirent blogs have highlighted some of the testing considerations for secure network slices, edge latency, and slice SLA management.
AI – the ‘I’ stands for infancy
AI and Generative AI (GenAI) were prominent topics at MWC Las Vegas and 2023’s answer to last year’s metaverse hype.
There was no shortage of pumped-up excitement around potential — details of actual progress were harder to come by.
AWS shared results of a recent survey indicating half of the 100 telecom executives surveyed expect to adopt GenAI within two years. Potential applications were highlighted in several keynotes, supplier-promoted platforms (IBM WatsonX), and products (Arrcus ACE-AI Networking Solution). There was also discussion and announcements on ecosystem partnerships (NTT Docomo & Nvidia).
Of course, it’s easy to talk. It was harder to find clarity on actual business cases and implementation roadmaps outside of GenAI’s role in customer service (chatbots) and marketing (personalization).
There’s clearly an important role for AI in networks for OSS/BSS automation, RAN optimization, and future self-driving and self-securing networks. But this all felt fragmented and lacked clarity on when and where AI would be used compared to today’s advanced analytics and early machine learning implementations.
Side conversations at the event tended to focus on the cost to train and build, the risk of poor and biased training data, and looming regulatory uncertainty.
AI and GenAI will no doubt play a role in the future of networks and services, but there’s work to be done to resolve data fidelity, privacy, and governance issues. This is an opportunity for test and assurance, which could help to:
Train with unbiased data and real-world feedback loops
Validate security efficacy
Provide governance in the form of assurance and visibility that the AI black box is delivering the relevant operational and business outcomes
On the operational side, the new demands of AI will have a large impact on data centers. They need to evolve rapidly to handle AI’s surging compute, storage, and especially networking requirements. The high speeds required will push the boundaries beyond 800GE, and lossless Ethernet performance will be critical. You can learn more about the impact of AI workloads on data center networks in our webinar.
Open RAN has blueprints, but concerns remain
The Open RAN market is beginning to coalesce around ecosystem requirements, and there has been progress on real, tested offerings. However, there’s a nagging concern about potential market fragmentation.
NTT DOCOMO continued to promote its OREX® Open RAN ecosystem packages that let service providers select from four disaggregated base station configurations comprising combinations of 13 ecosystem vendors. The packages have been selected, tested, and optimized for specific deployment requirements and are a positive step forward to demonstrate interoperability—but the result is decidedly less open than pure Open RAN.
Thorough testing of OREX packages is critical for service provider adoption. DOCOMO is using Spirent’s Open RAN capabilities for its 5G Open RAN Ecosystem (OREC) testing to test multi-vendor interoperability and real-world performance. According to Sadayuki Abeta, Corporate Evangelist at NTT DOCOMO, “Emulation is a critical element of any O-RAN testbed and Spirent is the market leader in this area. Including its technology as a key component of OREC testing will ensure operators can gain confidence in O-RAN deployments.”
On the transmission front, Ericsson’s preshow announcement highlighted an intent to integrate open fronthaul in its Cloud RAN portfolio in 2024, supporting Class A operation mode on RUs. Since vendors are free to select Class A or B, another interoperability paradigm raises its head. Around the show, Ericsson’s Class A decision prompted discussions about future fragmentation, cost, performance, and possibly vendor lock-in.
These creeping concerns could delay adoption.
Verizon, for example, said it is interested in “what Open RAN becomes,” but isn’t sure it can meet its current needs. The company noted key Open RAN challenges remain around interoperability, software, scalability, and performance.
So how should the Open RAN industry address these concerns?
In part, by building confidence through proof points. That is, demonstrate and test that vendor solutions are interoperable and real-world performance and resilience scenarios can be supported consistently. In addition, a continuous testing capability will be needed to revalidate efficacy as new vendors are swapped in and out, software upgrades occur, and underlying network changes happen.
Learn how Spirent's O-RAN test solutions are helping the Open RAN community on their journey to 5G Standalone.
Progress, but still challenges ahead
Whereas MWC Shanghai showcased regional advancement, MWC Las Vegas focused largely on opportunity and the hurdles that remain.
The progress in bringing suitable private 5G wireless solutions and Open RAN offerings to market is encouraging, however, there remains a long road ahead before widespread adoption. Testing will be essential to demonstrate viability as well as to assure ongoing performance.
AI, of course, presents enormous opportunity but remains in its infancy. It will be interesting to see what next year’s show brings regarding actual business cases.
For additional insights into the 5G story and to make sure you have the right testing strategy, read the eBook 5G Standalone and the Race for New Revenue.