Dish Network is preparing to build a 5G network that might allow it to circumvent many of the costs and challenges typically associated with operations and maintenance. However, its vision for a completely software-defined network using virtualized network functions and equipment and software from less traditional vendors is relatively unproven.

This is especially true on such a grand scale with a network that aims to reach all of the United States from sea to shining sea and almost every person, place, and thing in between.

Now that the Department of Justice, T-Mobile US, and Sprint have agreed to terms on a deal that will prop up and strengthen Dish Network to become a fourth nationwide operator, the hard work of building and operating a network still lies ahead. As a newcomer to the wireless industry it will encounter many challenges, critical decisions, and new levels of regulatory oversight. And that doesn’t even take into account how much this will cost (answer: as much as $20 billion, according to analysts).

Dizzying Array of Network Options

Every mobile network begins at towers or cell sites that power the radio access network (RAN). While this equipment and infrastructure is pretty well dominated by four vendors, Dish Network has hinted that it might get creative and look elsewhere.

Its choices for doing so are relatively limited. Dish has said it would like to emulate the strategy of Rakuten Mobile, which is designing a fully virtualized greenfield network in Japan, but RAN stalwart Nokia is providing some radios for the network. Rakuten Mobile is also working with NEC to jointly develop a 5G radio, but it will only support the 3.7 GHz spectrum band.

Dish could turn to U.S.-based companies Altiostar, Airspan, JMA Wireless, Mavenir, and Parallel Wireless, but “none of these vendors have a proven track record in terms of delivering a large scale radio access network,” said Daryl Schoolar, principal analyst at Ovum.

“You need experience. This is a big bet. If you’re going to spend $100,000 remodeling your house, are you going to hire some random guys who say they’re cheap or are you going to say ‘hey let me see some pictures of some work you’ve done?’” Schoolar said.

Dish Network could go with non-traditional vendors but “radio engineering is still more art than science,” said Chris Nicoll, principal analyst at ACG Research. “Antenna and radio is not where you skimp — thus Dish may turn to experienced vendors such as Airspan and others for metro RAN, but large scale macro RAN will likely remain with Kathrein, Nokia, Samsung, and Ericsson for radio and antennas,” he explained.

“The question is how bleeding edge do they want to be? Not using the traditional, proven vendors just for the sake of not using them does not make sense,” Nicoll said, adding that it’s still unknown if virtualized infrastructure vendors can scale their solutions to cover hundreds of millions of potential customers.

“When it comes to the supplier landscape for Dish, it is important to remember that Dish is not starting from scratch and the migration path from [mobile virtual network operator] to [mobile network operator] will vary somewhat with the different network functions,” said Stefan Pongratz, analyst at Dell’Oro Group. “For the low-band RAN network, clearly the larger RAN suppliers including Ericsson and Nokia are well positioned to provide a compelling [total cost of ownership] given their existing low-band footprint and their overall sub-6 GHz portfolios.”

Benefits of Virtualized, Software-Driven Network

In all likelihood, Dish Network will turn to a large incumbent for RAN equipment. But if it sticks to its stated goal of building a virtualized, software-driven network its choices broaden considerably.

“If they go down this route, some of the benefits should be more future proof because they do not have any legacy network to support,” said Schoolar. Down the line, even the oldest equipment in Dish’s network will be relatively new and flexible, and this would theoretically allow the company to mix up its vendors and turn up services quicker on a virtualized framework, he added.

Without a legacy network to support, Dish has an opportunity to deploy a virtualized, centralized network from the start, said Joe Madden, president at Mobile Experts. “The cost would theoretically be low for this kind of network due to the low hardware cost and the higher level of competition for software,” he explained. “This is unproven so far but in general it’s likely that the capex will be low. The remaining question is whether the operating cost will also be as low.”

Many analysts question how much network operators can save in capital expenses by taking this approach. Dish won’t be leapfrogging its competitors anytime soon, but the opportunity to offer competitive services is a definite possibility, according to Nicoll.

“It might allow [Dish] to build a better network that is ready for standalone 5G New Radio and network slicing before others,” said Anshel Sag, analyst at Moor Insights and Strategy.

Challenges With Dish’s Approach

Despite the rosy outlook, there are some drawbacks to this strategy because it’s so new and unproven at scale, said Schoolar. “There’s a sense also that it probably costs you more money on the network services and integration side. You have to spend more money there integrating more disparate vendors together.”

Gaining the ability to spin up and take down programs or applications as they are needed sounds good, but it isn’t working in a commercial mobile network yet. And even if it was, the benefit is hard to quantify, Nicoll explained. “The hype around this exceeds the reality at the moment,” he said.

Dish will also encounter a constant cycle of integrations, testing, ongoing software releases, and performance tuning on the network, according to Nicoll. “They better find an excellent systems integrator to help them along.”

Despite the low costs of a virtualized network, Dish will have to establish many new fixed assets including towers and backhaul, Madden explains. His research also indicates that the “capacity of a generic vRAN system will be lower than an optimized system from a top supplier such as Nokia or Ericsson. With lower capacity, Dish may not achieve the economics that they hope for,” he said.

“The virtual RAN (vRAN) approach will be good for coverage but will not achieve the high-density capacity of a fully optimized network. This will become a challenge in terms of cost per GB for Dish,” Madden added.

Lurking beneath the buzzwords and hype are real challenges that all of the more established and proven network operators are still trying to overcome.

“I would caution anybody if they start talking about this really advanced network — cloud, virtualized, you know a string of technology buzzwords,” Schoolar said. “Don’t be surprised if they kind of back off some of that as they go through the process because they may not find that everything out there is as ready as they think it is.”

Whereas T-Mobile US has earned widespread respect for its networking prowess and operations, Dish’s talent and ability as a mobile network builder and operator is a big unknown, said Avi Greengart, lead analyst at Techsponential. When T-Mobile US says it’s going to combine multiple networks and technologies as it merges operations with Sprint, it’s reasonable to take the company at its word, he said.

“When Dish says ‘we’ve never built a network before but we're going to build one using new vendors, technology that hasn’t even been approved yet, and it’s going to have additional functionality,’ well, first of all you haven’t even told me what the additional functionality is, and the execution risk is very high,” Greengart said. “Just saying it’s going to be software-defined network that uses a made-for 5G spectrum all dedicated to 5G — it’s an interesting idea, but I don’t know what it means.”