The cellular networks of the future will need the ability to quickly and intelligently adapt. As 5G technology gets closer to widespread reality, telecommmunications infrastructure will soon have to accomodate the complex connectivity and data demands of everything from self-driving cars to AR/VR experiences and a whole new generation of IoT devices.
Telco providers are keenly aware of the radical evolution that’s coming, which is why companies like Cisco are investing heavily in software-defined network automation. Ahead of Mobile World Congress, the networking giant has announced Cisco Crosswork, a new suite of network automation tools to imbue next-gen infrastructure with analytics, machine learning, and the ability to perform a host of proactive, data-driven operations.
“Customers’ appetite for bandwidth is insatiable, so we’re releasing more network elements to service that need,” said Jonathan Davidson, Sr. VP and GM of Cisco Service Provider Networking. “We’ve hit the tipping point where no single human can understand how the entire infrastructure is put together. We need to augment these capabilities with automation. Whether it’s planning or designing or implementing or optimizing infrastructure, you need software and machine learning to help abstract away the problems our customers are facing.”
Cisco Crosswork Network Automation consists of five different software offerings:
Crosswork Change Automation: An automated operations application for what Davidson called “closed-loop control,” meaning that when a problem happens on the network, it can be resolved without the need to engage a human. Instead, the software will analyze a problem, fix it, and notify the human it was fixed.
“You want humans writing software that talks to systems instead of humans talking directly to systems,” said Davidson. “We’re now incorporating data consolidation, machine learning, event correlation, and closed-loop change automation down to the individual network slice and eventually across multiple domains.”
Crosswork Health Insights: A suite of smart sensors, smart alerts, and smart remediation software for network monitoring and optimization. Davidson said these tools are part of a larger goal to build out self-healing infrastructure.
“Automating an upgrade on one of your devices the old way is manual using a script,” explained Davidson. “The newer way of doing it is a smart upgrade. Do a health check on the device and all the devices connected to it.”
Crosswork Data Platform and Network Insights: These tools, built on both an open-source and commercial data analytics platforms, are the key to what Davidson described as mass awareness: aggregating all the siloed data across a network and transforming it into something actionable. Then the cloud-based network insight solution then creates data visualizations to help solve large-scale routing issues.
“We’re unlocking the power of data by fitting all that data trapped inside various systems inside a centralized data lake and using augmented intelligence to put that data to use with analytics. We want to move to proactive control—a platform allowing you to take that augmented intelligence and use it to take control of infrastructure and solve problems before they become issues.”
Crosswork Situation Manager: This machine learning-based event correlation engine then takes all that data and uses it to generate select scenarios to help quickly remediate problems. Davidson explained that there’s also an added element of social interaction, with tools built into the products to help operators troubleshoot problems and share knowledge.
“The end goal is ultimately to get a self-healing and self-optimizing infrastructure up and running without service interruption or human intervention; customized learning that can evolve over time,” said Davidson. “You can also tie in your APM software like [Cisco-owned] AppDynamics and get alerts coming out of that tool to correlate together across multiple technology areas. All of this requres the ability to do machine learning and natural language processing [NLP] to train the machine and pull these things together.”
Ultimately, all of these automated tools come back to preparing network infrastructure for the imposing challenges on the horizion. Whether we see nationalized 5G infrastructure in the U.S. or whether the major carriers continue along their current pace, 5G will be here sooner rather than later.
Davidson said 5G is going to be a big driver of not only additional bandwidth, but new services that live a lot closer to where customers reside. Problems that once could be resolved deep in the network are now wrapped in new layers of complexity that begets new sets of issues to be solved. Cisco thinks the way to mitigate that is automation.
“The service providers themselves have been very aggressive especially in the last few months about how they want to get 5G infrastructure out in late 2018. There are vast amounts of land you need to cover, but I expect it to move quickly,” said Davidson. “There’s a little bit of conern about what new services will be enabled, but we’re starting to see the real potential of the new surfaces this [infrastructure] can generate, and I think it’s going to move faster than anyone thought a year or two ago.”