Hewlett Packard Enterprise (HPE) upgraded its HPE Aruba Networking Central platform for network management by integrating generative artificial intelligence (GenAI) large language models (LLMs). The upgrade enhances the platform’s search functionality, making it faster and more accurate, simplifying tasks for network operators, and enhancing network services.
HPE has already invested significantly in AI for network management, developing technologies for tasks like configuring environments and understanding device connections. According to Alan Ni, senior director of edge marketing at HPE Aruba Networking, the introduction of generative AI and LLMs further improves the platform’s search capabilities and enables more intelligent, human-like responses to queries.
The GenAI-powered search feature promises faster, more accurate, and more efficient query handling. This is achieved by replacing traditional natural language processing engines with more advanced LLM transformers, which better understand the intent behind queries and provide relevant responses.
“If you think about the queries that go into Central, they are network jargon. The efficacy of this is a lot better when we understand intent. We also can provide the end user with output based on that intent. It may be a direct command, where they’re instructing the system to understand why there is poor performance. So instead of giving a search result, it can bring you to a specific screen or data around that specific question,” said Ni.
Historically, troubleshooting networks required a high level of network knowledge to understand what data to look for, where to look, and how to correlate and analyze the information. GenAI democratizes access to information and enables users to ask, in natural language, questions about the network, resulting in better performance and faster troubleshooting.
HPE’s approach to AI development was to build custom models specifically for networking. Rather than relying on external generative AI services like ChatGPT, HPE developed its own models integrated into HPE Aruba Networking Central, running on the HPE GreenLake cloud platform. This approach not only enhances the precision of the search function but also provides better data protection. It prioritizes privacy and security, operating within a secure environment where the LLMs are self-contained.
HPE’s extensive data lakes are fueled by approximately four million network devices and over a billion endpoints. By collecting data from network devices and customer interactions, HPE can improve its predictive capabilities and offer more accurate suggestions using AI. This approach enhances the AI features and minimizes the risks associated with inaccurate or unsuitable responses, according to David Hughes, chief product officer at HPE Aruba Networking.
“It’s not burdened by a whole lot of data that isn’t relevant for the IT admin, but it’s specifically being trained on the exact data that will be helpful. By narrowing the scope of the information that can be regenerated, it’s much more useful and safer,” said Hughes.
Hughes emphasized that HPE does more than just networking. Delivering AI, security, and automation through HPE Aruba Networking Central puts the company at the forefront of “tearing down barriers” and allows network administrators to do their jobs without requiring additional training in this technology.
Many IT vendors have relied on off-the-shelf models, which are typically much cheaper than building them in-house. The purpose-built models, specifically designed for the network, should result in more accurate results. HPE also creates its silicon for the same reasons. Off-the-shelf processors are good at many things, but custom ASICs are great at networking. In networking, particularly in today's digital world, those slight differences can significantly impact how a network runs.
Since its launch in 2014, HPE Aruba Networking Central has offered network management tools for setting up, managing, monitoring, and fixing both wired and wireless networks. The platform operates as a cloud-based service with two subscription levels, Foundation and Advanced. The new AI-powered search engine will be accessible to all subscribers in the second quarter of fiscal year 2024. It can also be accessed as an independent service and part of the HPE GreenLake for Networking subscription.
Zeus Kerravala is the founder and principal analyst with ZK Research.
Read his other Network Computing articles here.
Related articles:
Modernization of the data domain to include new practices, tools, and technologies is the third phase of digital transformation.
Think of Hypershield as an AI-enabled security fabric where every network port becomes a high-performance enforcement point.
With the rise of AI and its massive impact on networking, Cisco has pivoted swiftly to enable customers to utilize its converged networks to run taxing AI apps.
