Be afraid. Be very afraid. In the not-too-distant future, a few giant Internet firms with multi-trillion-dollar valuations will shape public opinion, cater to every need and have the means to switch off our most critical infrastructure.
AT&T's decision to entrust the most intelligent part of its 5G network to Microsoft, as reported by Light Reading yesterday, gives the software company the power of a brain surgeon over his patient. One slip of the scalpel and the lights may go out.
There is much that is counterintuitive about these public cloud deals, and this one especially.
Telecom operators balk at dependence on a cabal of massive equipment suppliers, as shown by the enthusiasm for open RAN, an attempt to inject competition into the mobile infrastructure market.
Yet AT&T is apparently happy to let Microsoft be the sole custodian of its 5G core. If there is an alternative, it is not mentioned in AT&T's statement.
Likewise, operators insist they want a bigger say over the development of technology. It explains why they are so active in groups like the Telecom Infra Project (TIP), and bristle at suggestions TIP is "Facebook-led."
It is why Vodafone CEO Nick Read says he wants to "insource IT development engineering capability" and develop his own intellectual property (IP) "to make our differentiation stronger."
But AT&T's latest deal entails a transfer of its engineering and lifecycle management software, along with hundreds of employees, to Microsoft, which will package them into the Azure for Operators service it sells to other telcos. It is a download of expertise into the world's second most valuable company (behind Apple), stripping AT&T of IP it must once have considered important.
Resistance is futile
So why do it? Largely because the attractions of the public cloud were too great for a shrinking, debt-burdened and strategically inept telco to resist.
The widespread view is that operators can save money by shifting IT workloads and network functions into AWS, Microsoft Azure or Google Cloud, the power brokers of today's Internet. They can also cruise in the supercar of public cloud innovation and forget about their own sputtering engines.
Dish would also have been a catalyst. The satellite company is building a fourth mobile network in the US and has gone even further than AT&T, recently deciding to park its entire IT and software estate inside AWS.
"For us to be able to create more value for customers and for private networks, it is important to be as flexible, software-wise, as we can," Marc Rouanne, Dish's chief network officer, previously told Light Reading.
"If you can move your observation data very easily, and put it in data lakes and do machine-learning correlation, then you can learn a lot of things about the optimization of networks."
AT&T's copycat move does certainly not preclude differentiation, according to James Crawshaw, a principal analyst with Omdia (a sister company to Light Reading), or mean the operator has relinquished all engineering assets.
"It is a bold but pragmatic move," he says. "There are still plenty of mobile network engineering activities that AT&T will retain and these still give it plenty of scope to differentiate from its competitors. They seem to be taking a view that running a telco cloud is not a source of differentiation and best left to someone with greater scale, like Microsoft."
But a fitter company that had ceded less ground to big tech would probably not have countenanced a deal. AT&T's share price has lost a third of its value in the last five years after a disastrous foray into media, marked by its $85 billion splurge on Time Warner.
Its overall debts stood at $157 billion last year, up from just $66 billion a decade earlier. In 2020, it recorded a $3.8 billion loss. Its workforce has shrunk by 50,000 employees since 2015. It has been overtaken by T-Mobile, a seemingly unstoppable competitor. Dish threatens a price war.
No doubt, the undisclosed commercial arrangements with Microsoft will be economically favorable, but they seem to completely ignore the dangers of single-supplier deals.
Those have been starkly illustrated in the telecom sector by recent sanctions against Huawei, forcing some telco customers of the Chinese equipment maker to switch vendors at considerable cost. Nokia's earlier dependence on Intel for 5G chips is another cautionary tale. When Intel ran into manufacturing delays, Nokia's share price tanked.
Ditching and switching
What happens if AT&T simply decides it would like to switch vendor, or even take systems back in-house? Others' experience of using the public cloud suggests it would be tricky. In 2019, Snap, the provider of the Snapchat messaging service, warned in a regulatory filing that moving systems between AWS and Google Cloud would be "difficult to implement" and cause it to incur "significant time and expense."
US authorities have also raised the alarm. In its October antitrust investigation, the House of Representatives said customers trying to move between public clouds had encountered "high switching costs" and "technical design challenges."
In a paper about using the public cloud for operational support systems, Crawshaw advised telcos to avoid overreliance by splitting workloads between providers and keeping 10% of them in a private cloud.
Not everyone is persuaded by the economic arguments in favor of the public cloud, either. Andreesen Horowitz, a venture capital firm, thinks cost savings are usually ephemeral. In a recent report about the public cloud, it said "the pressure it puts on margins can start to outweigh the benefits as a company scales and growth slows."
It also drew attention to the difficulty of repatriating systems. Most companies, it said, "find it hard to justify moving workloads off the cloud given the magnitude of such efforts."
The latest, telecom-specific snub has come from Tareq Amin, the high-profile technology executive in charge of building a mobile network for Rakuten in Japan. AWS and other public clouds are "years behind" his own company in running cloud mobile networks, and any notion they can do it at lower cost is "completely wrong," he recently told reporters.
Self-serving as those comments are – Rakuten, after all, aims to sell its cloud services to other telcos – they should make operators think twice about following AT&T.
As the US communications company settles into its new Microsoft surroundings, it needs to hope for a couple of things: that more attractive alternatives to Microsoft do not materialize; and that its new long-term partner does not turn abusive.
- AT&T to offload 5G into Microsoft's cloud
- AWS 'many years' behind in cloud mobile, says Rakuten CTO
- MWC's big story is still the cloud cheerleader replacing Ericsson
- Fastly outage gives preview of Internet apocalypse
- Bell Canada is latest telco to succumb to AWS at the edge
— Iain Morris, International Editor, Light Reading