Every new Wi-Fi generation arrives with the same promise: more speed, more capacity, more spectrum. Wi-Fi 6 brought efficiency to crowded buildings, Wi-Fi 6E opened the 6 GHz band, and Wi-Fi 7 added multi-link operation for higher throughput. Buyers notice the numbers on the box, and the market rewards them. The global enterprise wireless LAN business was worth roughly $25 billion in 2025 and is on course to reach about $45 billion by 2030. Yet the specification printed on a datasheet is not the product an enterprise actually deploys. What arrives in a hospital, a warehouse, or a campus network is software, and software is where most launches quietly succeed or fail.
Shalabh Parmar has spent more than 13 years on the side of that problem a datasheet never shows. A Senior Staff Engineer at Qualcomm, he leads product engineering and quality work for enterprise, retail, and carrier Wi-Fi, and he has helped some of the industry’s largest vendors move new access points from lab validation into live customer networks. He is also a named inventor on a U.S. patent for increasing capacity in multi-link Wi-Fi operation, the same feature buyers now see advertised on Wi-Fi 7 hardware. His argument is plain: the reliability of an enterprise network is decided long before the box is opened, in the quality of the software at the source.
Speed Gets the Headlines. Software Decides the Outcome.
The visible story of enterprise Wi-Fi is a story of standards. Wi-Fi 6E extended networks into the 6 GHz band, and Wi-Fi 7 introduced multi-link operation, which lets a single device use several bands at once. Each step adds throughput, and each step also adds firmware, drivers, and control logic that must behave correctly across thousands of client devices from hundreds of manufacturers. The hardware sets the ceiling. The software decides whether a network reaches it, or spends its first months in the field chasing defects that testing never caught.
Parmar saw this early. From 2019 to 2021 he led the chipmaker’s engineering integration and issue resolution for one of the first enterprise Wi-Fi 6 access point platforms to reach the market, built by the networking vendor Aruba. He took on a product engineering role that made him the technical bridge between the silicon supplier and the customer, and he ran the resolution of more than 200 software defects, including 20 serious enough to block the launch outright. He coordinated firmware, physical-layer, and driver teams, ran daily reviews, and stood up focused engineering groups to close the hardest problems before shipment.
“A datasheet tells you what the silicon can do in ideal conditions,” Parmar says. “A customer network tells you what actually happens when hundreds of device types, old and new, associate at the same time. The gap between those two is software, and closing it is engineering work, not marketing.”
Enterprises Cannot Absorb Late-Cycle Surprises.
For a business, a wireless network is now infrastructure in the plainest sense. When it degrades, work stops. More than 90% of midsize and large enterprises estimate that a single hour of downtime costs them more than $300,000, and in regulated sectors the figure climbs much higher. A software defect that surfaces after deployment is not a line in a bug tracker to the customer. It is a clinic that cannot pull records, a distribution center that cannot scan pallets, a trading floor that drops connections during market hours.
Parmar’s answer to that pressure was to move the point of detection earlier. Rather than treat quality as a final inspection, he built practices that surfaced customer-facing problems during development, when they were cheaper and safer to fix. He represented the customer’s interests inside the company’s engineering reviews, so real field priorities shaped what got fixed first. He set up periodic software baselines that were shared with customers and run through regression testing, so every change could be traced and validated instead of bundled into a risky late-stage release. The aim was fewer surprises reaching the customer, not more heroics once they arrived.
“Late fixes feel productive because everyone is busy,” Parmar notes. “They are actually the most expensive way to build anything. Every problem you catch at the source is one the customer never has to live through, and reliability is really just the sum of those choices.”
Reviewing the Field Sharpens the Work.
Enterprise wireless has grown harder to reason about. A modern campus carries laptops, phones, sensors, cameras, scanners, and building systems on the same network, each with its own quirks. New spectrum and new features widen the testing surface with every generation. Engineers who want to build dependable products have to stay close to how the whole field is moving, not only their own corner of it.
Parmar keeps that perspective in part by evaluating other people’s work. He is a peer reviewer for IEEE Access, the engineering journal, where assessing new research keeps him current on methods beyond any single product cycle. That habit feeds back into his day job, where he has led quality across four Wi-Fi generations, from Wi-Fi 6 and 6E through Wi-Fi 7, and now drives the quality effort for the emerging Wi-Fi 8 standard. Each generation repeats a familiar lesson: the newest capability is the least tested, and the least tested is where reliability breaks first.
“Reviewing forces you to argue with an approach you did not invent,” Parmar observes. “You learn to spot the assumption that will not survive contact with a real deployment. After enough of that, you start catching the same assumptions in your own designs much earlier.”
The Cost of Skipping Root Cause.
When enterprise systems go dark, the network is the usual place to start looking. Networking and connectivity problems are the single most common trigger of IT service outages, ahead of software, power, and cooling, at 31% of incidents. Many of those failures trace back not to a dramatic event but to a defect that shipped, sat quietly, and surfaced under load. Treating the symptom restores service for a while. Only finding the root cause keeps it from returning.
Parmar has built much of his practice around that distinction. He treats serviceability, meaning how quickly a problem can be diagnosed in the field, as a design requirement rather than an afterthought, and several of his patent applications come from inventing better ways to observe and resolve issues in live networks. He also reviews for ACM’s Computer-Supported Cooperative Work and Social Computing conference, part of a wider habit of studying how complex systems, and the people who run them, actually behave. The through-line is a preference for understanding a failure completely over clearing it quickly.
“Closing a ticket and fixing a problem are not the same thing,” Parmar explains. “If you do not know why something failed, you have just scheduled the next outage. Serviceability is what lets you get to the why while the network is still running.”
Designing Reliability In, Not Inspecting It After.
The next wave of enterprise wireless will lean on multi-link operation, wider 6 GHz channels, mesh coordination, and early edge AI to manage congestion. Every one of those capabilities adds power, and every one adds new ways to fail. The pattern does not change with the standard. Capability arrives first, and reliability has to be built to match it.
Parmar’s position is that the industry already knows how to close that gap and often chooses not to. Quality treated as a checkpoint at the end of development will always trail the product. Quality built into the lifecycle, with customer issue patterns steering engineering priorities, honest early-warning signals, real release gates, and serviceability tools, moves reliability upstream to where it is cheapest to secure. The enterprises deploying this infrastructure do not measure success in peak throughput. They measure it in the disruptions that never happened.
“Faster hardware will keep coming, and that is a good thing,” Parmar reflects. “But the customer never experiences a specification. They experience the software on top of it, on a bad day, under full load. Build for that day at the source, and the rest of the promise takes care of itself.”
