Amazon Web Services (AWS) and Dish have announced that Dish will be using Amazon Elastic Compute Cloud (EC2) to host its RAN and mobile core for both public and private 5G networks. This marks the first time we're seeing a major telco — the 4th largest mobile network operator in the U.S. — bet its operational network on public cloud infrastructure.
In an analyst briefing this week, Dish and AWS shared that Dish will use public cloud resources in AWS Regions and AWS Local Zones and private deployments of AWS Outposts in Dish's network locations and on customer premises to build its 5G network. Marc Rouanne, Dish’s chief network officer, indicated that the company had spent 18 months evaluating multiple infrastructure options, including other public cloud providers and cloud-native options, before settling on AWS.
AWS-Dish partnership details
Dish and AWS are testing 5G RAN on AWS infrastructure, with 5G RAN DUs running on AWS Outposts and 5G RAN CUs hosted in AWS Local Zones. Other 5G core and IMS components will be hosted in AWS Regions (although I expect some of these components will run in Local Zones or AWS Outposts in Dish networks).
AWS Local Zones are cloud resources owned and operated by AWS located in major metropolitan areas. Today, Local Zones are available in LA (with two availability zones), preview versions in Boston, Miami, and Houston, and a total of 15 sites committed by the end of this year (additional sites in Atlanta, Chicago, Dallas, Denver, Kansas City, Las Vegas, Minneapolis, New York, Philadelphia, Phoenix, Portland, and Seattle). Dish's intention is to leverage these sites to run critical parts of its RAN, in addition to thousands of AWS Outposts located in Dish's network. The company intends to use AWS's deep AI expertise, cloud-based orchestration and leverage cloud development CI/CD pipelines to help accelerate its own pace of development.
As part of Dish's deployment commitments to the FCC, the company will need to build a network that covers 20% of the U.S. population by June 2022 and extend that to cover 70% a year later. Dish believes that making a bold move of leveraging a public cloud, like AWS, will help it achieve its goals.
Dish and AWS indicated they have successful proof of concepts today and intend to launch this into production later this year. While not an exclusive relationship, Dave Brown (VP Amazon EC2, AWS), Stephen Bye (Dish chief commercial officer) and Rouanne all intend to keep the architecture simple — which I read as keeping the number of involved parties small.
Beyond the RAN
Beyond serving Dish's RAN and mobile core needs, the company expects to use AWS for other IT workloads and as a host for private enterprise networks. AWS is touting its Arm-based Graviton2 CPU and Nitro architecture to run telco workloads more efficiently and securely. Both parties hope to tap into the AWS developer ecosystem and extend EC2 APIs into the network to allow AWS partners to tap into 5G networks services via new APIs. Notable is Dish's dependence on AWS' backbone to transfer data from the Local Zones into the AWS Regions.
RELATED: Chua: Edge computing in 2021 —When elephants dance
AWS and Dish both shared that Dish's selection of vendors (many picked before Dish made the AWS decision) have been verified to run on this infrastructure. AWS points out that many of those are AWS telco infrastructure partners. Dish had previously announced Allot, Nokia, and Palo Alto Networks as its security vendors, Mavenir and Altiostar for RAN, Netcracker for automation and orchestration, Amdocs for billing and VMware for its virtualization platform. These are part of 35 announced vendors in contract across radios, tower, backhaul network core, and security.
What the AWS-Dish partnership means
At AvidThink, we've been tracking telco-hyperscaler partnerships and announcements for a while now —from IT to IoT, OSS to BSS, edge computing to private networks. While we've seen the use of AWS (and Azure) private infrastructure to host private 4G/5G networks, this announcement is a breakthrough in the potential scope and scale. Carriers have been happy to leverage cloud economics for IT and back office functions, and those relationships abound. Telcos have been skeptical of moving their operational components for production networks onto the public cloud — until now.
This bold decision leaves Dish dependent on AWS for the performance, availability, security, and resiliency of its RAN and core. Dish SLAs will be a joint responsibility between both parties. We view this as a bold move, not dipping a toe or a foot into the water, but leaping in and making a splash. More than a technology or business decision, this sets the tone within Dish for going cloud-native. We've worked with many enterprises and telcos in the past, and we see a marked difference between organizations that embrace CI/CD, cloud-native, as-a-service thinking, and those that are burdened by legacy culture. Embracing a cloud-native platform can benefit innovation and agility, encouraging employees to think and build things flexibly and tap into the larger developer ecosystem. It could help recruitment, as Rakuten has shared about its cloud culture versus a telco culture.
The announcement is fresh, and there are plenty of details to be worked out. Areas that we'll be closely watching include:
- Impact on the VMware-Dish relationship — Previously, Dish had announced using VMware as its infrastructure and application orchestration. Dish shared that it will likely use VMware for specific workloads (e.g., VM-based), but from our standpoint, if you're betting on a public cloud partner, it's more efficient to go with native orchestration and automation. Putting an abstraction layer in place increases portability and allows Dish to go multi-cloud later, but we're not confident that the overhead is necessary. With the proper automation framework, Dish could adapt to another cloud without a virtualization/abstraction layer.
- The orchestration tango — With VMware and Tanzu, Netcracker, AWS, it'll be interesting to see how they all come together. Rouanne indicated a multi-layer orchestration strategy, which is pretty standard in terms of the usual business/intent, services, domain, infrastructure, and element orchestration stack. However, the hyperscale cloud providers all have their opinions on how this should be done (e.g., Azure Arc, Google Anthos) at the infrastructure level, and all have sights set up the stack.
- SLA management, 5G network API development — A strong SLA for 5G and network APIs to unlock 5G value will be necessary for telcos and hyperscalers. Not only at the network edge but on-premises for private networks. For us, we'll be interested in whether a telco built on a public cloud platform can demonstrate end-to-end SLA and 5G API development faster than the others. How SLA violations and uptime issues will be resolved between the two parties will be worth tracking.
- Costs and economics for Dish — The financial details of the AWS-Dish tie-up were not shared. We have to expect significant negotiation and discounting — Dish isn't going to be paying those regular hourly EC2 on-demand charges or the robust monthly AWS Outposts rental fees. How using a public cloud provider impacts Dish CapEx, and OpEx will be an exercise for the financial analysts.
- Mobile edge rollout — With the Dish RAN components hosted on AWS Outposts and AWS Local Zones, it's a simple matter, as AWS's Brown hinted, to treat those Local Zones also as mobile edge sites. It's not clear what impact this will have on the Wavelength brand and the Verizon relationship. Eventually, all major carriers will have edge computing relationships with all hyperscalers, so I'm sure it will not come as a surprise. This does make running workloads in Local Zones appealing since those sites can serve both wireline and wireless end-users with low latency (perhaps not as good as Wavelength).
- Unlocking other carriers for AWS — Now that Dish has made the leap, will other carriers be more open to using AWS to host their operational network infrastructure for wireline and wireless networks?
The AWS-Dish announcement is significant, and we are looking forward to seeing it in production and rolling out across the nation over the next few years. And we won't be surprised to see other tie-ups, like, say, a Dish-powered Amazon mobile service. Alexa, Call Me Maybe?
Roy Chua is founder and principal at AvidThink, an independent research and advisory service formed in 2018 out of SDxCentral's research group. Prior to co-founding SDxCentral and running its research and product teams, Chua was a management consultant working with both Fortune 500 and startup technology companies on go-to-market and product consulting. As an early proponent of the software-defined infrastructure movement, Chua is a frequent speaker at technology events in the telco and cloud space and a regular contributor to major leading online publications. A graduate of UC Berkeley's electrical engineering and computer science program and MIT's Sloan School of Business, Chua has 20+ years of experience in telco and enterprise cloud computing, networking and security, including founding several Silicon Valley startups. He can be reached at [email protected]; follow him at @avidthink and @wireroy
Industry Voices are opinion columns written by outside contributors—often industry experts or analysts—who are invited to the conversation by FierceTelecom staff. They do not represent the opinions of FierceTelecom.