Self-Healing Networks Help Healthcare Organizations Optimize Performance and Enhance Security

How Networks Falter, Then Heal Themselves

One of the biggest responsibilities of network administrators is maintaining network uptime. When an outage occurs, troubleshooting the problem can be a time-consuming process that negatively affects patient outcomes or experiences that depend on the network. Clinicians and hospital administrators and contractors may be unable to fulfill their duties, which can lead to major impacts to patient care and trust in the organization.

Self-healing networks offer an optimized approach to network management that uses artificial intelligence and machine learning (ML) to continuously monitor network performance, analyze traffic patterns, detect potential issues, and take corrective action directly or through administrators to maintain network stability and avoid downtime.

“The need for healing can come from a number of different perspectives; it can be an equipment malfunction, poor deployment, a misconfiguration or changes in network requirements based on organizational changes,” says Larry Lunetta, vice president of AI, security and networking product marketing at HPE Aruba Networking. “And it can also be that the network has been compromised. There could be malicious activity that changes the network in a way that requires remediation.”

“A self-healing network leverages AI-native automation to maintain optimized performance,” says Christian Gilby, senior director of product marketing for AI-native networking at Juniper Networks. “The network identifies the issue, looks at the data and then figures out what’s wrong, and depending on the circumstances either remediates the issue or tells the admin how to resolve it.”

READ MORE: How can healthcare organizations create self-aware and secure IT networks?

What Defines a Self-Healing Network?

The growth of AI and ML technologies is crucial to the development of this networking innovation, but a wider array of technologies is needed for a network to be truly self-healing.

“In addition to capabilities like artificial intelligence and machine learning, you need toolsets that can leverage automation and an infrastructure that can be subject to automation workflows,” says Felipe Fernandez, CTO at Fortinet Federal.

What makes a network self-healing can vary, but it requires networking components including switches, routers, access points and cables that are built to support and respond to the monitoring, analysis, automation and optimization being overseen by the AI and ML solutions. This can also include a software-defined WAN solution to enable a more secure and flexible virtualized network architecture.

Data lakes are typically needed to store all of the data being collected and analyzed, as intense data analysis informs these processes. Given the overall complexity of the network and self-healing processes, generative AI assistants can provide an easy interface for network administrators to interact with to oversee activity on the network.

Optimizing Operations With Self-Healing Networks

One of the primary benefits of a self-healing network is that it can automate delivery of a reliable user experience, especially when dealing with changing network conditions.

Physicians and clinicians working in the field, such as performing surgery in another country or during an emergency situation, must quickly stand up Wi-Fi resources. With many variables affecting Wi-Fi signals, it can be a challenge to deliver a reliable care delivery system.

“Our HPE Aruba Networking Central console has an AI insight that will make recommendations on how to change the settings in the Wi-Fi to make sure you deliver a good customer experience,” Lunetta says. “But there’s no compromise of the performance from passerby traffic inadvertently connecting to the network.”

Another benefit of a self-healing capability is being able to boost overall IT efficiency and reduce the more time-consuming responsibilities of network administration.

“One of the benefits is eliminating some of the manual effort that consumes a network administrator’s day,” Lunetta says. “Think about how much time is taken out of their day for data gathering, then correlation, then analysis, before finally making a decision to fix an issue.”

DISCOVER: These are the top five healthcare challenges solved by observability tools.

The Security Potential of Self-Healing Networks and Automation

Along with the ability to optimize network operations comes the security value that self-healing networks can deliver.

“There are a range of security events and attacks where a self-healing network can help with recovery: unauthorized access, DDoS attacks, malware propagation,” Fernandez says. “Because these networks are tied into network performance and understanding traffic flows and security event logs, they can take that data and apply automated remediation or mitigation actions that can respond to threats that are occurring and bring network services back online.”

Self-healing networks present a powerful way to automate security responses, but figuring out how far to automate security and operations responses remains a challenge.

“From a security standpoint, self-healing networks best manifest themselves at the gateway; either you get on or you don’t with a certain set of permissions,” Lunetta says. “After that, I think there are a lot of judgments that people still have to make. You can implement different actions based on the severity of the alert and the confidence you have, but this gets to the issue of trust: How much are you willing to automate security responses?”

RELATED: Understand the differences between observability and monitoring in healthcare.

As the AI managing the network has more time to monitor network activity, it gradually develops a baseline understanding of typical traffic and user behaviors. This allows self-healing networks to improve over time at getting ahead of potential operational and security problems.

“There’s a lot of hybrid use today with self-healing networks,” Gilby says. “In many cases, it’s identifying the issue, but then there’s still some manual intervention. It’s not just about if the technology can do it and do it accurately; it’s also a trust thing. IT teams need to be at a point where they can fully trust it.”

The level of automation that network administrators are comfortable with will depend on the healthcare organization’s culture and the level of trust it is able or willing to hand over to the network itself.

“Every organization needs to look at the issue of trust and apply their own metrics or logic on the governance. They need to apply a level of trust based on the level of autonomy they want the system to have, and there are workflows within self-healing networks that allow for human intervention,” Fernandez says. “When certain conditions arise, you can bring a human into the loop to apply their intuition and knowledge of external factors to dictate what that next action is going to be.”