Your browser is not supported. Please update it.

12 October 2020

Dish partners with Ciena for slicing; Ergen casts his wireless net wider

Dish Network continues to build up its supplier list, adding Ciena’s Blue Planet platform for network automation. It continues to boast that it will have a fully virtualized, multivendor and open 5G network, Rakuten-style (though with different suppliers) next year, despite some analysts’ scepticism about Dish’s commercial chance of making a big dent in the market share of the big three MNOs, especially as it is also trying to prop up an ailing pay-TV business.

But Dish’s chairman, Charlie Ergen, is never one to be deterred by a risky commercial landscape, and is now looking beyond Dish’s 5G build-out to more ambitious investments in wireless technology. He plans to create a company – separate from Dish but with the ability to support it in technology terms – which would focus on acquisitions and roll-ups in the wireless industry. He hopes to raise about $1bn for this effort.

His plan has echoes of Softbank’s Masayoshi Son and the Japanese maverick’s plan to use the massive Vision Fund to support every key element of digital technology. That is in tatters now, and has forced Softbank to sell most of its stake in Sprint and so exit the US market – the event which propelled Dish into 5G with a deal to buy Sprint’s Boost Mobile prepaid unit, as a sop to competition regulators.

Ergen disclosed his plan in a regulatory filing this month, detailing a ‘blank check’ company called CONX which would raise money via an IPO to “effect a merger, capital stock exchange, asset acquisition, stock purchase, reorganization or similar business combination with one or more businesses or assets”. In its initial securities offering, CONX intends to offer at least 100m units priced at $10 each, aiming to raise a total of at least $1bn.

This kind of ‘special purpose acquisition company’ (SPAC) is commonplace in the USA, with over 100 launched this year, and are designed to marry up private companies with respected management teams, with investment, in a faster way than via IPO.

Potential targets will be selected based on “compelling long term growth prospects”, “leadership in technology-driven transformation”, “low or manageable risks of technological obsolescence” and a “defensible position in intellectual property”, says the filing.

No targets have yet been specified but of course, investors’ eyes are on some of the suppliers to Dish’s 5G network, which include Altiostar and Mavenir on the RAN software side (the latter has just filed for its own IPO).

Meanwhile, Marc Rouanne, EVP and chief network officer at Dish, provided the latest progress update at VMware’s virtual conference last week. VMware is providing the telco cloud platform for Dish’s 5G build-out, supporting a 5G Standalone virtualized core from Nokia plus the RAN software, and working with Ciena for automation and Fujitsu radios.

Rouanne was keen to downplay the fears of many operators, that there will be unacceptable performance trade-offs when deploying a virtualized RAN rather than an integrated one, given the hugely demanding nature of many RAN tasks, especially when moving towards low latency 5G.

“A lot of people think open RAN is challenged when it comes to performance, but because you have access to all the data and because you have access to much more computing power, all of a sudden you can push the limits of spectrum and capacity much further,” Rouanne said, adding that Dish can now onboard software at the radio and cloud-native level in hours or, at worst, days, rather than weeks or even months for traditional systems.

“When you have cloud-native capabilities, it’s very easy to change some microservices or some capabilities without impacting the whole chain,” he added. One of VMware’s functions in the Dish roll-out is to certify and verify all the vendors’ cloud-native functions and how they work together.

VMware recently repackaged its MNO offerings under the banner of 5G Telco Cloud Platform. This supports a full software stack from hypervisor and networking, up to automation, Kubernetes orchestration and network function orchestration.

“When you design a completely dynamic network, you don’t decide where things are going to be upfront. You let the network design itself. So we have access to all these stacks and capabilities on VMware, and the orchestrator will push them where needed, dynamically,” Rouanne said. This orchestrator will enable “one of the first large-scale distributed data lakes”, which will extract, analyze and store information at the edge of the network, moving out as far as edge nodes on aggregation sites or even cell sites. The edge cloud, which mirrors a similar approach by Rakuten, will use components from multiple vendors.

Rouanne added: “5G is a more cloud-driven and software-driven business than radio-driven, and that’s something that is not well understood in the marketplace. As you go down the mast, now you hit the cloud and that’s where it’s completely different and it looks like mini data centers, mini private networks, or big data centers if you go all the way up.”

One of the end results of all this greenfield experimentation should be an early ability to support end-to-end network slicing, which remains challenging to implement in full for operators with significant legacy equipment still in place. “Everything we’re designing now is geared to slicing,” said Rouanne.

That will help the company to support many different customers with specific connectivity and QoS requirements, and is likely to go far beyond Dish’s traditional pay-TV base. While adding its own mobile strand to its existing services to revive their appeal with quad play bundles will be foundational to the business case, it will not be enough to justify years of investment in spectrum and a major build-out, given that Dish’s core satellite TV business is in decline. Even if the ultimate aim is to prove the value of the spectrum and sell it on, which would fit the Ergen playbook, there will need to be a more convincing commercial case.

That would logically lie in the enterprise market, where the early activity in the CBRS shared and licensed bands has indicated the amount of pent-up demand for connectivity which enterprises, cities and other groups can control themselves, and which is tailored to their specific requirements. Supporting such demand via a wholesale business – with industrial MVNOs taking a dynamic slice to support their enterprise customers – would fill a gap which has existed for years in the USA. Google and Amazon have talked about virtualized wholesale networks to support enterprise MVNOs on a dynamic, pay-per-usage basis; LightSquared (now Ligado) had the same vision but has been stymied by regulatory issues. Dish could sell direct to enterprises, but that would require a very different approach to business, so a wholesale system – providing higher QoS and controllability for enterprise service providers than unlicensed or shared spectrum – would be a quicker route to market.

A critical enabler of slicing, along with the cloud-native 5G core, is full automation on every layer, so that the virtual networks can be created and optimized on the fly, pulling together the appropriate network resources and numbers of instances to suit a particular use case. Ciena’s Blue Planet software for automation, inventory management and service order fulfilment will be central to the slicing plans, said Rouanne.

He explained: “In a traditional network your inventory is the sum of many parts built over time. It’s hard to have one global inventory of the network, then it’s very hard to slice it and automate it because you don’t have a real time database that tells you at any time how your network is doing.”

Blue Planet’s software provides the inventory database, detailing all the capabilities of the network at any given time from the core to the edge, and then matches those resources with customer orders that have specific requirements such as bandwidth, coverage and latency.

Blue Planet has been building up its slicing proposition over the past couple of years, acquiring three companies to boost its closed-loop automation – DonRiver for real time inventory management; Packet Design for multilayer assurance; and Centina for service assurance management.

But like Rakuten (see Open RAN Watch), the challenge will be to translate architectural innovation and efficiency to commercial success in a mature and competitive market. Wall Street analyst firm MoffettNathanson Research has recently published sceptical reports about the commercial challenges both for Rakuten in Japan and Dish. For the latter, they recently reiterated a ‘sell’ rating on the shares, with a target price of $15 per share.

With no clarity on wholesale plans, the analysts believe the current presumed net present value of the wireless business, at $15bn, is “irrationally optimistic”. Dish’s spectrum is worth more than that, they say, but only if the network is built out and proven, which will introduce many costs that will count against the current NPV. The analysts have repeatedly questioned Dish’s target cost for its 5G build-out, set at $10bn, questioning whether this includes costs of labor, tower leases, backhaul and other outlays which are only marginally affected by virtualization.

“Our bearish view of Dish is not based on scepticism about Dish itself so much as it is our scepticism about the wireless market,” wrote analyst Craig Moffett. “In the end, our valuation and target price for Dish rest on a simple conceptual question. What is the likely NPV of a greenfield entry into the wireless business in the United States? If one assumes, as we do, that the debt on Dish’s satellite TV business is likely already greater than the enterprise value of that business, but that the debt on the satellite business does not have recourse to the yet-to-be-built wireless network, then the equity value of Dish shares is simply equal to the NPV of whatever it is that Dish decides to build.”

Dish has to build out the spectrum because of FCC obligations imposed with the Boost deal with T-Mobile. If Dish cannot provide 5G services to at least 20% of the US population by June 2022, and hasn’t yet deployed a core network supporting services in its AWS-4, Lower 700 MHz E Block and AWS H Block spectrum, Dish will be fined $16m for each frequency band that misses the population goal by 25% or less, $32m for each band that misses the goal by 25-50%, $48m where the miss is 50-75%, and $66m for more than 75%.

By June 2023, if Dish is unable to cover at least 70% of the US population with the same bands, it will pay $6m for each percent missed by the AWS-4 band, $2m in the 700 MHz E Block, and $2m in the AWS H Block band.