Various big names not so familiar in the cellular industry are starting to make their presence felt as 5G gets rolling, determined to seize a major stake in what they see as a field now well suited to their products and technology. They even threaten the incumbent players, or at least would like to, as Oracle made clear recently in stating it planned to challenge the dominance of the mobile core by the big three – Ericsson, Huawei and Nokia – which account for around 80% of that sector between them.
Some also want to go beyond the core and deliver operators their whole 5G infrastructures via cloud-based SaaS (software-as-a-service) models. This comes at time of major disruption anyway in the cellular ecosystem, given the furore over Huawei’s involvement in the network core in many US-aligned markets, which threatens to partition the whole supply chain into two distinct camps.
Apart from Oracle, we see Microsoft, IBM, Dell, HPE and Cisco all raising their game substantially, in some cases having been plotting and quietly expanding their portfolios over a decade ever since 5G first appeared on the horizon.
There is a long history of major enterprise IT players entering the legacy fixed telco arena. Cisco, after winning the IP router wars in the early to mid-1990s, to become the dominant supplier of such technology in the enterprise sector, extended first into the then-burgeoning Internet and from there into the telco space. But it never fully made the transition into mobile during the years of major expansion in the late 1990s and 2000s, which has led to the situation where the USA, despite its dominance of many IT tech sectors, has been left without a major all-round force in cellular infrastructure, certainly beyond the core. As a result, US carriers have become reliant on the big two – since Huawei has been excluded from national networks for a decade.
But now that could start to change. Indeed, these major US-based IT players will be encouraged to invest more heavily in the cellular field by the country’s determination to become more self-sufficient in 5G. For instance, Oracle, which has built out from early dominance in database technology, has just announced plans to enter the field for 4G and 5G core network services this year, offering them as-a-service from its cloud.
The company believes this will appeal to operators because it will convert capex into a recurring opex with costs matching revenues, avoiding need for over-provisioning. Of course, such promises can be a little disingenuous because Oracle itself will have to indulge in some over-provisioning to meet peak demand, until it develops the scale to smooth these out across its global infrastructure. Until then, these costs will still be born indirectly by operators, even if they still benefit from paying as they go.
But Oracle is able to boast that, having been in the fixed telco space for many years as well as having a large established cloud, it will quickly reach that required scale. It also has an integrated internal supply chain, even down to its own chipsets. And Oracle has been extending its tentacles into the actual telco network for some time, having acquired two vendors of IP equipment in 2013 – Tekelec for an undisclosed sum and session border control vendors Acme Packet for $1.7bn in 2013.
HPE and Microsoft are two other majors to make big plays this year, motivated again by the opportunity afforded by the 5G transition to seize some share from those big three in the core. A key point that helps explain the timing is that, while early 5G deployments will be based on slightly enhanced 4G cores via the NSA (Non-Standalone) model, the 5G SA (Standalone) specifications are complete. A few operators are ready to start migrating to, or adding, 5G SA networks, which require a full 5G core, often in cloud-native mode.
The big enterprise players want to be there from the outset. HPE announced in March that it would have a pre-integrated core package available in the second half of the year, allowing operators to cherry-pick the components they want, or have them all slotting together. These cloud-hosted components include Unstructured Data Storage Function (UDSF), Unified Data Repository (UDR), Authentication Server Function (AUSF), Unified Data Management (UDM), and Network Exposure Function (NEF).
Meanwhile, Microsoft appears to be following even more closely in Oracle’s footsteps as it belatedly attempts to position itself for, it hopes, a 5G gold rush. It has splashed out a similar amount on acquiring NFV capability by buying Affirmed Networks in March 2020 for about $1.35bn, followed by Metaswitch Networks for an undisclosed amount. The company has been edging into enterprise telecom services for some time but now, like Oracle, plans to exploit its huge cloud, Azure, to offer network services to enterprises, and potentially carriers as well, either fixed or mobile.
Cisco is a rather different case, having been a provider of infrastructure equipment from the outset. It is curious that it missed the opportunity to become one of the major players in cellular on the back of its dominance in IP routing and switching. However, Cisco did make an early move towards a mobile core that could be deployed separately from the RAN, when in 2009 it paid $2.9bn for Starent, a start-up specializing in separating the packet core. Cisco followed up in 2011 by helping some members of the Starent to create a new start-up, Altiostar Networks, which offers virtualized RAN software that runs on (relatively) commoditized hardware.
Cisco has, so far, been unsuccessful when it tried to address the RAN directly and move away from its enterprise roots. While its company and public WiFi business is strong, cellular products acquired with companies like Navini and Ubiquisys were shelved, and now Cisco is selling the SON (self-optimizing network) it acquired with Intucell in 2013 (see separate item).
The router giant appears to have been waiting for 5G to generate more serious commercial opportunities and had several announcements planned for Mobile World Congress 2020. Two of these were offerings designed in turn to help mobile operators derive new enterprise revenues from 5G services as well as LTE.
One was Cisco Cloud Services Stack for Mobility, a platform for operators to bring new LTE and 5G services to market, backed up by the requisite computational and network management tools for a virtualized environment. The other was LTE FWA (Fixed Wireless Access) Managed Services, which Cisco pointed out had already been taken up by Verizon and AT&T as part of their plan to deliver FWA services to enterprises and, in Verizon’s case, consumer households.
While some of these IT majors make a virtue out of being one-stop shops for the core at least, others trumpet the value of partnerships to amass complete services for operators. In May 2020, IBM raised its stake in 5G by announcing various initiatives and partnerships focused primarily on network edge computing. This was significant in being the first major expression of IBM’s acquisition for $34bn of RedHat to bring in the world’s largest agglomeration of open source software and expertise. Closed in July 2019, this takes IBM right back to the top table of IT, having survived more disruptions of technology than any other vendor as its erstwhile competitors of yesterday have mainly fallen by the wayside and forgotten (with the obvious exception of HPE).
We should also not forget Amazon and Google, although their ambitions increasingly look broader – to take on the carriers themselves, even if that may involve partnerships with operators or acquisitions. Indeed, Amazon has already partnered with four telcos – KDDI, Vodafone, Verizon and SK Telecom – for its AWS Wavelength, a platform for developers to build edge applications. AWS compute and storage components are embedded within those telcos’ data centers at the edge of their 5G networks, which can then access the full range of AWS services in that area.
This may seem like a good model for the operators to gain those capabilities, but also extends Amazon’s tentacles ever deeper into 5G. The message seems to be that 5G is too important to be left to the operators and their traditional technology suppliers, as was largely the case for 4G.