Red Hat is a company that provides telecom infrastructure software. It’s probably most associated with its work providing container-based software through its OpenShift operating system. But in addition to providing software for telecom core networks, it also provides software for the radio access network (RAN).
Recently, the company has been working to improve energy efficiency in both the RAN and the core of wireless networks.
Rimma Iontel, chief architect of the global telco team at Red Hat, is focusing on sustainability and energy efficiency.
She said ever since telecom networks have begun disaggregating RAN components into the centralized unit (CU) and distributed unit (DU), more software has been introduced, and Red Hat has been involved.
Red Hat works with different partners in the RAN space — including Nokia, Ericsson, Samsung and Mavenir — to optimize the underlying server hardware. For instance, the partners might tweak how much the CPU core consumes energy. “If you break into containers, you can identify parts of the DU that don’t require such top-level performance, can tweak the frequency as to when the CPU runs and the settings when the CPU goes idle to save power,” she said.
Iontel said Red Hat is also working with Juniper Networks in the open RAN space to use Juniper’s RAN intelligent controller (RIC) to optimize energy efficiency.
Of course, Red Hat, also helps telcos save energy in their data centers.
RELATED: Nokia leans on Red Hat for open 5G core
For instance, Red Hat works to improve the energy efficiency of the Intel servers that run its OpenShift software. One way it does that is by optimizing the placement of workloads. “There’s a sweet spot for each server,” said Iontel. “There’s a certain cutoff where it cannot consume less power without actually turning it off. And a certain point where you cannot optimize any better. There is this sweet spot of energy use versus node use and the performance of what’s running on it where they intersect.”
Iontel said the various techniques to improve energy efficiency are not yet done dynamically, although that would be the goal in the future — to leverage artificial intelligence for dynamic optimization. Currently, energy usage is optimized by understanding how the equipment, the software and the applications and services work.
She said AI for energy efficiency “is a little bit more of a future type of thing. Right now we know the behavior of the application, so we prepare for that. With AI there are not enough models yet to really do in production at scale everywhere.”
Red Hat’s work versus other energy managers
There are companies, such as Galooli, that do energy management for wireless networks at a different level. For instance, Galooli helps tower companies and operators to optimize energy usage at tower sites. Galooli’s software monitors and recommends energy allotments between grid electricity, batteries, diesel generators and renewables.
Iontel said Red Hat is interested in bigger-picture energy usage of the type that Galooli manages.
She said there is a lot of legacy equipment in telco networks and no data on how much energy each individual component is consuming. “What we’re trying to do in cloud deployments is understanding where the energy is being consumed. Ideally, all of these aspects are going to come together to give us the biggest savings.”
Currently, Red Hat and its partners are creating tools to collect metrics to continue learning and improving energy efficiency. Iontel said Red Hat’s energy efficiency work is all open source software and is contributed back to the Cloud Native Computing Foundation in the Kepler project.