As the number of connected medical devices continue to rise, and as artificial intelligence (AI) continues to influence the broader healthcare ecosystem, network infrastructure is becoming the “nervous system” of an organisation’s operations that ensures the movement of data is both efficient and secure.
According to Vikas Butaney, senior vice president and general manager of Cisco’s secure routing and industrial internet of things (IoT) division, the increase in connected medical devices and AI applications in healthcare is creating the need for new network architecture that is purpose-built for AI.
Go deeper with GlobalData
Discover B2B Marketing That Performs
Combine business intelligence and editorial excellence to reach engaged professionals across 36 leading media platforms.
“The network needs to accommodate higher data volumes, expanded bandwidth requirements, and ever lower latency needs,” Butaney told Medical Device Network.
“As data moves between devices, applications, and vendors, it must be securely transmitted and handled in compliance with regulatory requirements. The network is evolving into more than just a conduit for data – it’s becoming the nervous system of an organisation’s operations.”
With health data being some of the most private and highly regulated information, secure data handling is non-negotiable.
Butaney continued: “Organisations must manage who has access to the data, where it’s stored, and how it’s used.
US Tariffs are shifting - will you react or anticipate?
Don’t let policy changes catch you off guard. Stay proactive with real-time data and expert analysis.
By GlobalData“A robust network can support these requirements by embedding security at every layer, ensuring compliance with regulations like the Health Insurance Portability and Accountability Act (HIPAA) while maintaining operational efficiency.”
According to Butaney, the future of real-time data processing for medical devices will also be highly dependent on advancements in AI, modern network infrastructure, and edge computing.
Edge computing refers to distributed computing models that bring data processing closer to the source of the data, as opposed to relying on centralised cloud or data centres.
“As more devices generate and process data in real time, the ability to analyse that data at the edge will become critical for reducing latency, optimising bandwidth, and delivering actionable insights at machine speed,” Butaney shared.
Evolving cyberthreats as connectivity and AI deployment continue to rise
As AI-driven workloads and data traffic volumes rise due to connected medical devices and the deployment of AI models in healthcare systems, the attack surface is expanding.
According to GlobalData analysis, the AI market in healthcare is projected to reach a valuation of around $19bn by 2027, while healthcare providers’ security cybersecurity spend is projected to grow at a compound annual growth rate (CAGR) of 12.5% to a valuation of $10.9bn by 2027.
As AI-workloads rise across healthcare, Butaney foresees a rise in “vibe coders” that may presage an increase in “vibe hackers” who use generative AI (genAI) tools to identify and exploit vulnerabilities.
“This evolving threat landscape demands a fundamental shift in how we approach security, with protection embedded into every layer of the network,” Butaney said.
“By fusing security into medical devices, data transport, and device identity, organisations can scale real-time workloads confidently, ensuring both resilience and reliability in an AI-powered world.”
As AI becomes more prevalent in healthcare, the limitations of HIPAA are becoming increasingly apparent. In today’s highly technological landscape, updating the 1996 regulation is becoming more essential than ever.
Aaron T. Maguregui, partner at law firm Foley & Lardner recently told Medical Device Network: “HIPAA was intended to scale with time and with technology. What I don’t think HIPAA ever contemplated was the fact that AI would be able to essentially take in data from multiple sources, match it together, and create the potential for the reidentification of data that was never intended to be used for reidentification.”
