In 2001, standing before the IEEE INFOCOM crowd, Mike Hluchyj described network convergence as a 25-year march. The journey had begun for him as a graduate student, when a single network for data, voice and video still belonged partly to theory. It followed him into research laboratories, carrier switches and a startup attempting something that sounded reckless to the telephone establishment: move voice onto packets and trust the architecture. His keynote looked back at what had changed and ahead to what the next quarter-century might bring. Twenty-five years later, the answer is sitting in plain sight. Networks swallowed nearly everything. They also became difficult to see, govern and defend. Hluchyj’s latest company lives inside that consequence.
Uptycs is usually described with the busy vocabulary of cloud security: telemetry, workloads, endpoints, posture, detection. Strip away the labels and its proposition is remarkably familiar. Gather what a complex system is doing. Give the activity a common structure. Let people ask precise questions of it. The nouns have changed since Hluchyj studied packet traffic. The intellectual habit has not.
01 / Origin packetThe future began as a traffic problem
Hluchyj earned his electrical-engineering degree from the University of Massachusetts Amherst in 1976, then completed master’s and doctoral work in electrical engineering at MIT. He joined the technical staff at AT&T Bell Laboratories and later became director of networking research at Motorola Codex. These were useful places for a person interested in queues, switches and the awkward fact that every network looks generous until everybody tries to use it at once.
His technical reputation formed around the theory, design and analysis of high-performance packet networks, work that led to his election as an IEEE Fellow. Yet the career did not remain inside papers and laboratories. By 1994, he was vice president and chief technology officer at Summa Four, a maker of programmable switches for carrier networks. Research had become product architecture. The controlled problem had acquired customers, budgets and competitors.
02 / SonusA telephone call becomes software
In 1997, Hluchyj co-founded Sonus Networks. He served briefly as its founding president, then as CTO, vice president and secretary. The company entered a telecom world built around circuit switching and argued for an all-packet future. Voice had exacting demands. Delay is an inconvenience while loading a web page; in conversation it becomes interruption, overlap and annoyance. Sonus needed to make internet telephony behave like infrastructure rather than an experiment.
Hluchyj later explained that the company had rejected a parallel time-division backplane and committed to packets. The team squeezed gateway delay to roughly two and a half milliseconds, low enough that early prospects reportedly doubted the number. It separated routing logic, subscriber data and applications into distinct elements. When standards caught up, the architecture already looked familiar. “We were the pioneers in developing SIP as the interface from the softswitch to the application server,” he said in 2005.
Sonus went public in 2000. The following year, Hluchyj’s INFOCOM keynote gave the bet a longer historical frame. He was not selling a miraculous break with the past. He was describing the market’s belated arrival at an idea researchers had been worrying over for decades. That distinction matters. Technology fashion likes a lightning bolt. Infrastructure usually arrives by committee, iteration and somebody patiently measuring the delay.
One question, several networks
03 / The London tellThe poker face and the bad morning
The most revealing public anecdote about Hluchyj involves an ambush by reality. In London in 2005, a reporter arrived at an early meeting carrying fresh news: BT had named suppliers for its enormous next-generation network, and Marconi, Sonus’s partner, had been left out. The public-relations minders visibly registered the shock. Hluchyj stared into the distance with the composure, the reporter wrote, of a poker professional considering an audacious bet on a terrible hand.
He eventually admitted surprise. Then he said Marconi remained a good partner, noted other opportunities and left room for the future. It was a founder’s answer in the best and most maddening sense: acknowledge the card, continue playing the table.
His more durable poker strategy was engineering vigilance. He said Sonus trained its hardware teams to ask two questions: “How could someone else do this better? How could somebody beat us?” Cost pressure began with architecture, not panic. Components were negotiated hard. Old patterns were treated with suspicion. His compact instruction, “Start with the architecture,” could sit over nearly every subsequent turn in his career.
04 / The second tableVerivue, Akamai and a team that returned
Hluchyj left Sonus in 2006 and co-founded Verivue, serving as CTO. The network problem had moved again. Video and other content were multiplying; operators needed delivery infrastructure closer to their own networks. Verivue built licensed content-delivery systems for them. In December 2012, Akamai bought the company for $30.9 million in cash, seeking to accelerate its operator CDN business. Hluchyj became CTO of carrier products and worked on service-provider initiatives involving network-functions virtualization.
The transaction was an exit, but the more consequential asset may have been a working relationship. Ganesh Pai, Uma Reddy and Hluchyj had accumulated years together across Sonus, Verivue and Akamai. When they started Uptycs in 2016, they were not a group assembled by a networking dinner. They had already watched one another make decisions when the architecture was unclear and the news was bad.
That history gave Uptycs a peculiar advantage. Security startups must ask customers to trust software with a panoramic view of their machines. The founders, meanwhile, had already tested their own trust through multiple technical eras. Repeat founders are often praised for pattern recognition. The quieter benefit is that nobody needs a rehearsal for candor.
05 / Query the machineCloud security, in rows and columns
Uptycs began with an open-source tool called osquery, which presents operating-system information as database tables. Processes, users, packages and connections can be questioned with SQL. For a security team drowning in specialized consoles, the appeal is almost comic in its plainness: ask the computers what they are doing in a language many analysts already know.
The company paired that collection method with cloud analytics, correlating activity across fleets of laptops, servers and workloads. The product expanded into the broader terrain of hybrid-cloud protection, but the original idea still contains Hluchyj’s signature. Separate collection from analysis. Normalize the moving parts. Plan for hundreds of thousands of endpoints, not the handsome laboratory demo. Make the system legible enough that an operator can reason about it.
His current title is founder and board member, which places him away from daily product narration and closer to the long view. That suits a career measured in transitions. Circuit voice gave way to IP. Content delivery moved into distributed infrastructure. Servers dissolved into cloud workloads and containers. Each shift created abundance, then complexity, then a market for somebody who could see the system whole.
A patent count that kept moving
06 / The campus loopEngineering meets another discipline
There is a second long collaboration in Hluchyj’s story. He and Theresa “Terry” Murphy met as UMass Amherst undergraduates, he in engineering and she in nursing. In 2008 they established a graduate fellowship for students interested in collaborative research. In 2021 they provided the initial funding for the university’s Center for Nursing and Engineering Innovation, designed to bring the two disciplines into the same room before habits harden.
The choice feels consistent with his technical career. An interface is where separate systems become useful to one another, and also where misunderstandings breed. Nurses understand daily practice. Engineers understand how to formalize and build. Put them together early enough and each may prevent the other from solving an imaginary problem with great efficiency.
In 2026, UMass Amherst gave Mike and Terry Distinguished Achievement Awards for their support of interdisciplinary innovation. He also serves on the engineering college’s dean’s advisory council. It is an elegant return: the student who began following packets now helps create a place for other kinds of knowledge to cross.
07 / The through-lineArchitecture is memory with a plan
A résumé separates Hluchyj’s life into employers. His work makes more sense as a single diagram. Bell Labs and Motorola supplied the grammar of networks. Summa Four made switching commercial. Sonus carried conversation across packets. Verivue organized the delivery of content. Akamai widened the operating scale. Uptycs turned the behavior of modern machines into something security teams could query.
Across the diagram, the recurring questions are practical and slightly severe. Where should intelligence live? Which functions should remain separate? What breaks when ten thousand becomes ten million? How much delay will a user tolerate? What can the operator actually observe? And, lurking behind all of them, how might a competitor design the whole thing better?
The technology business adores reinvention, a word that often means forgetting the old vocabulary. Hluchyj’s career offers the opposite pleasure. It shows technical memory accumulating rather than being discarded. A packet researcher can recognize a cloud-security problem because both require choices about traffic, visibility, scale and control. The future keeps changing its interface. Underneath, the old questions wait patiently.
In 2001, Hluchyj wondered what another 25 years of convergence would bring. The answer was not a single triumphant network. It was billions of connected systems producing more evidence than any person could inspect. Uptycs is one response: collect the evidence, give it structure, ask better questions. Somewhere inside the cloud, the packets are still moving. Mike Hluchyj is still following.