The shift to cloud, SaaS, and hybrid work is no longer breaking news. What is surprising? How many IT and network teams are still trying to stitch together architectures that weren’t designed for today’s distributed world. Data is everywhere. Users are everywhere. Applications live across SaaS, public cloud, and private data centers. Yet too often, traditional network and security architectures can’t keep up, creating bottlenecks, security gaps, and user frustration. Modern SASE and SD-WAN architectures promise agility, resilience, and better performance, but only when they’re…
AI infrastructure is growing at an unprecedented pace. Enterprises are racing to build clusters of GPUs, scale up AI workloads, and modernize their data pipelines. Yet one critical layer is often overlooked in these initiatives: the network. While AI and data leaders focus on compute, storage, and models, the network quietly becomes the bottleneck. Traditional, static networks—built for legacy application traffic—can’t handle the dynamic, latency-sensitive, high-throughput demands of distributed AI workloads. And without visibility, orchestration, and automation across the full stack, enterprise IT leaders are…
Lightyear’s new report highlights telecom industry and pricing trends observed in its proprietary dataset, illustrating how these factors may impact enterprise IT infrastructure strategies. Lightyear, the leading provider of enterprise telecom management software, release its highly anticipated 2025 State of Connectivity Report. This comprehensive report utilizes Lightyear’s proprietary dataset to provide timely insights on telecom industry and pricing trends. “We at Lightyear are rethinking the enterprise telecom experience from first principles, and a significant component of doing so involves helping enterprises better inform telecom decisions with…
As technology progresses and permeates all facets of modern life, the sophistication and scale of cyber threats continue to grow, presenting a formidable challenge to businesses worldwide. The consequences of a data breach can be catastrophic, meaning cybersecurity is at the top of the priority list. As such, it’s imperative for organizations to find efficient and effective solutions to protect their networks, safeguard sensitive data, and ensure business continuity. Cybersecurity Challenges in Modern Networks With your watch, refrigerator, and car all connected online, malicious actors…
Why has network automation failed to live up to expectations? After more than two decades of effort, millions of man-hours, and hundreds of tools created, network automation has largely not achieved its intended objective of eliminating manual operations. Multiple industry association surveys (e.g., Enterprise Management Association, Gartner) show that most enterprises have automated less than half of their data center tasks. Why is this so? When asked, most IT network operators would say that “automation” should help them do manual, repetitive tasks faster. But fast…
In today’s dynamic network environments, traditional MELT telemetry (metrics, events, logs, and traces) falls short when uncovering “unknown unknowns”—unforeseen issues that can cripple network performance and endanger network security. Packets and metadata generated from the packets data offer a more granular, real-time view, providing the level of detail necessary to identify these undetected anomalies and veiled vulnerabilities. The Shortcomings of MELT Telemetry While MELT telemetry can alert you to “known unknowns”—anticipated issues you’ve prepared for—its reliance on aggregation and predefined triggers leaves it blind to unexpected network events or…
On Oct 23-24 in New York City, we’ll be at the ONUG AI Networking Summit Fall 2024. Join us at the show and talk to Antonio Nearly, an LLM-powered holographic recreation of HPE CEO Antonio Neri that showcases an array of cutting-edge AI technologies. Antonio Nearly is highly knowledgeable about everything HPE, and you can ask him about anything else too — from professional soccer teams to blue whales. All you have to do is press a button and speak your question into a mic. It works…
Out-of-band monitoring techniques are necessary for AI clusters to provide trustworthy inferences. Out-of-band solutions provide latency analysis, decrease points of failure, and do not add additional burden on the network. [1] For AI clusters, the result of high latency are erroneous inferences. AI clusters are nodal network of GPUs that store and process inferences from machine learning models. [2] A slow latency for AI clusters inference is enough to produce incorrect inferences. The nodes of an AI cluster can create incomplete calculations on requests due to…
I recently came across an insightful blog by Bruce Davie in The Register that sheds light on the advancements in virtual network automation. While his analysis of virtual networks is insightful, he raises an important question about the state of physical network automation, “Are we actually any closer to the automation of networking than we were a decade ago?” My answer is a definitive ‘Yes’. Back in 2014, when network automation for physical networks was still in its infancy, we faced a steep learning curve….
The expanding domain of Generative AI is quickly becoming a pivotal aspect of enterprise innovation, with an impressive 66% of participants in IDC’s Future Enterprise Resiliency and Spending Survey recognizing its critical role. As this landscape evolves, the management of AI workloads across diverse environments becomes a crucial challenge, with issues such as latency, security, and connectivity at the forefront. In our view, a hybrid and multicloud networking strategy is not merely an option but a necessity for enterprises that leverage AI and operate within…
