29 Sep Microsoft raises the stakes in telecom fight against AWS and Google | Light Reading
A graphic in Microsoft’s latest blog about its 5G ambitions neatly encapsulates the Windows inventor’s telecom sprawl.
Its long fingers now reach into nearly every corner of the market, beckoning communications service providers (CSPs) to enter its embrace.
Cloud solutions? Goes without saying. Network functions? Check, after this year’s takeovers of Affirmed Networks and Metaswitch. Need an edge platform? Microsoft can bring that by simply extending the Azure cloud into new locations.
It even has an interconnect offer to advertise.
“They seem to think they’ve got it all covered,” says James Crawshaw, a principal analyst with Omdia.
Microsoft is not the only Internet giant on the telecom prowl. Both Amazon (through AWS) and Google have been stalking the CSP neighborhood, luring operators into cloudy partnerships.
Some telcos have joined hands with a hyperscaler to serve enterprise customers. Others have been convinced to run their back-office and IT services from a public cloud.
But neither Amazon nor Google has gone as far as Microsoft.
“This is the boldest, most coherent pitch to telecom by a public cloud player to date, by a distance,” says Gabriel Brown, a principal analyst with Heavy Reading.
“AWS and Google are in the mix but haven’t communicated anything like this level of ambition.”
Going after the network
What really distinguishes Microsoft from its two rivals is the former’s keen interest in network functions, says Brown.
“They are really targeting the carrier network itself.”
An operator could, in theory, run a packet core developed by Affirmed Networks entirely through the Azure public cloud.
It would be kissing off a critical part of its network, entrusting its security to Azure. But Microsoft could be a better home for it in straitened times, especially with the demands of 5G.
No one has done it, though, and many would be horrified even to consider awarding custody of network functions to Microsoft.
For that reason, Microsoft is desperate to look placatory.
“Recognizing that not everything will move to the public cloud, we will meet operators where they are whether at the enterprise edge, the network edge, or in the cloud,” it says in what Brown describes as a “cleverly worded” blog.
For the software giant, the immediate opportunity probably lies in enterprise and the Internet of Things.
“Where it makes sense is in private networks,” says Crawshaw.
A telco might have little interest in expanding its data center to accommodate smaller firms with their own private networks, he explains.
“That business might only last a year, so you host the packet core for those customers in Azure.”
The risk in this enterprise sector is that Microsoft becomes a telco rival instead of a telco partner.
Armed with Affirmed, Metaswitch and any other network suppliers it acquires, Microsoft might see little need for telco involvement, especially if a customer has its own spectrum.
The CSP could be left with the role of dumb pipe, providing basic connectivity a fate many are determined to avoid.
But it may be too late to start worrying about Microsoft as an enterprise competitor.
“That horse has already bolted,” says Brown. “Microsoft is an enterprise company it is not coming onto turf it did not already own and cloud connectivity has already taken over from WANs [wide area networks].”
Crawshaw also thinks the concern is somewhat overblown.
“Telcos are good channel partners and have field technicians who can get antennas pointing in the right direction,” he says.
“Microsoft won’t look to completely disintermediate telcos. They will just look to suck up the high-value bits and leave telcos to do the grunt work.”
Only the brave
Once operators have seen Azure-hosted functions at play in the private network sector, some brave or destitute telco will eventually decide a main core network that lives in the private cloud is simply too much trouble.
At that point, it outsources management, depreciates its server equipment and fires 90% of its data center staff, says Crawshaw.
If the move translates into beefier margins, others could flock to the public cloud like Walmart shoppers on Black Friday.
But who would take the plunge?
“At the moment, and probably for the foreseeable future, most operators of reasonable size are not going to want to put core network functions on a third-party public cloud,” says Brown.
“In some cases, there is visceral opposition to doing that.”
The likeliest candidate, in his view, is a mobile virtual network operator, or a telco entering a new geographical region.
Until then, one area of the CSP market remains a Microsoft-free zone, according to the software giant’s graphic.
Active post-takeover in packet core, and with its own lineup of voice and other applications, Microsoft still has no role in the radio access network (RAN).
But Rakuten, the Internet firm building a fourth mobile network in Japan, owns a majority stake in Altiostar, which develops RAN software. And it has now bundled Altiostar into its own telco cloud platform and made this available to other firms.
Could Microsoft buy a rival such as Mavenir or Parallel Wireless and do likewise?
It is feasible but unlikely, says Brown.
“I can’t see them wanting to get into the classic mobile RAN infrastructure market. There is too much hard work and headcount and too many processes they are not familiar with.”
Crawshaw agrees the hardware hassle would be off-putting.
“Obviously they don’t want to get into the antennas and all that gubbins, but the software part could make sense.”
In its blog, Microsoft does call out the radio access network as an example of a function it could support not necessarily through an acquisition but conceivably in partnership with a specialist such as Samsung.
A takeover, though, would make its graphic look complete.
Iain Morris, International Editor, Light Reading