Dish Network Corp. has put Kafka at the foundation of its new 5G Smart Network.
The Dish 5G Smart Network is an open platform on which developers can run applications. The ability for applications to get information from the network about performance, availability and routing is a critical feature, which is in part enabled by Dish's use of Apache Kafka vendor Confluent and its Confluent Cloud platform running on AWS. Apache Kafka's open source event data streaming technology helps to connect one source of data to another.
"One of our objectives is to build our network in a completely autonomous way, which will involve the use of intelligent functions performing different operations in a coordinated manner," said Brian Mengwasser, head of marketplace and app design at Dish Network. "As a critical part of our data infrastructure, we think of Confluent, and Apache Kafka in particular, as a central nervous system that enables coordination of different functions."
Event data streaming with Confluent supports 5G apps
Mengwasser noted that Dish needs to have the ability for different components of its network to react to situations as they arise. For example, if there is a performance issue, there can be a need to optimize data transport or to adjust wireless radios for 5G to help address a customer-facing issue. Beyond just providing information about the network, he noted that Kafka can also be used to expose new information and features to applications running on the smart 5G network.
"Kafka is one of the ways that we coordinate and ensure that the relevant information gets where it needs to be in a highly efficient manner," Mengwasser said.
With its smart 5G network, Dish is building a platform for enterprise applications somewhat analogous to how a smartphone is a platform for mobile apps, Mengwasser said. An enterprise can use the Dish Network Development Kit, talking to a network API, to develop a new application or extend an existing application into the network, he added.
"We want the network to serve the application and not have the network be some sort of black box that may or may not fulfill the enterprise's requirements," Mengwasser said. "The ability to customize the network to serve the application and connect with Kafka is what we're unlocking, and we think that's truly disruptive."
With Kafka, Dish is able to distribute intelligence and data around the network, rather than just dumping telemetry information into a single location like a data lake, Mengwasser said.
"It's a critical part of our data platform that we don't have to centralize because we can operate on data where it resides," Mengwasser said.
Dish didn't want to build and support Kafka on its own, which is why it chose Confluent, the lead commercial vendor behind Kafka.
"It makes much more sense to work with experts rather than reinventing the wheel," he said.
Dish also said this week it will work with Verica Continuous Verification Platform for Kubernetes and Kafka. Verica uses a chaos engineering approach to simulate challenges, allowing Dish to identify and proactively fix software issues before customers experience connectivity disruption, the company said.