EUC World Amplify 2026 Day Two: Skipping Forklift Migrations, DEX Without Alert Fatigue, Entra Travel Policies, and Running LLMs on Private Infrastructure

extract

EUC World Amplify 2026 continued Wednesday, September 30, at Milwaukee’s Baird Center, with a second day of technical sessions, architecture panels, vendor demos, and the hallway conversations that are the real reason most people come. After spending Tuesday on cloud migration strategy, Azure Virtual Desktop, endpoint identity baselines, and an interview with Nerdio CPTO Scott Manchester, we came into Wednesday with a wider view of how production shops manage, secure, and tune EUC infrastructure at scale. Figure 1: Baird Center entrance marquee for EUC World Amplify 2026. Before the expo floor got crowded, we walked it to talk with vendors. The mix covered enterprise mainstays like HPE and Nutanix, desktop management platforms like Nerdio, and startups including Tassient and Nevona.ai, alongside Citrix, Omnissa, Parallels, Liquidware, Leostream, IGEL, 10ZiG, Numecent, Recast, and Scylos. Figure 2: Some of the vendors at Amplify. Morning Kick-Off and Community Honors The day opened with a welcome from the World of End User Computing (WEUC) Board of Directors. Building on the independent community roots we tracked in our event preview, the board walked through the transition from the legacy Citrix User Group Community (CUGC) and other vendor-specific conferences into the broader, multi-platform World of EUC foundation. Figure 3: Wednesday’s speaker roster. The board then presented Johnny Ma with the WEUC Community Award, formerly the Stephanie Roper Award, an annual honor for an outstanding member of the EUC community that recognizes peer mentorship and technical leadership. We caught up with Johnny on the expo floor to congratulate him. Figure 4: Congratulating Johnny Ma on the WEUC Community Award. What sets Amplify apart from vendor-run user conferences is its focus on what practitioners deal with over polished marketing decks. Between a panel on Digital Employee Experience (DEX) monitoring and an open discussion of engineer burnout, the agenda was stacked with architecture deep dives: large-scale performance tuning, EUC policy automation, and diagnosing client-side compute bottlenecks. Figure 5: Wednesday agenda in the Baird Center Grand Ballroom. Independent vendors such as Leostream and Parallels showed broker orchestration and remote-delivery alternatives, which lines up with our ongoing coverage of VDI architectures and endpoints. As IT organizations work through Windows 11 hardware cutovers, shifting multi-cloud licensing, and workstation AI requirements, a vendor-neutral forum where administrators compare notes on session latency, broker failover, and policy automation earns its keep. DEX Panel: Telemetry vs. Alert Fatigue The technical program opened with a panel on Digital Employee Experience (DEX) with Vivek KR of Citrix, John Yoakum of Recast Software, Doug Arcidino of Omnissa, and Jack Smith of Liquidware, moderated by Sean Massey. Figure 6: DEX Panel with Citrix, Recast, Omnissa, and Liquidware. Having tracked DEX since its early days, we’ve watched it go from niche to essential. For years, IT teams monitored end-user computing by checking hypervisor CPU utilization, memory allocation, and ICMP ping. Across remote home offices, edge thin clients, and multi-cloud fabrics, those telemetry points miss what the user is seeing. The panel agreed that DEX pays off only when administrators look past dashboards and heat maps. Modern telemetry turns anecdotal support complaints into measurable root causes: an endpoint security agent eating 50 percent of client threads, a regional ISP transit hop masquerading as virtual desktop lag, or an oversized cloud instance with an expensive vGPU sitting idle. The bigger opportunity, the panel said, is feeding DEX telemetry straight into service management platforms like ServiceNow so remediation starts before a helpdesk ticket lands after a laptop battery degrades or an application crashes. The panelists were cautious about automating anomaly detection and summarization to relieve support desks, and we agree: unattended remediation scripts need strict guardrails and synthetic verification loops before anyone lets them loose on production. The App That Can’t Die: 2,000+ Concurrent Users on End-of-Life Stacks The Parallels Premier Sponsor session, titled “The App That Can’t Die: Serving 2,000+ Concurrent Users on End-of-Life Everything,” took on a problem nearly every systems architect eventually faces: keeping a critical line-of-business application running when the operating systems and frameworks under it are years past vendor support. The case study was a multi-tenant Parallels Remote Application Server (RAS) environment delivering Skyward’s K-12 school management software to more than 2,000 concurrent users across hundreds of districts, isolating the unsupported legacy backend from outside network exposure while keeping administrative control. An MSP architect running private cloud infrastructure for thousands of seats offered the session’s most pointed takeaway: a multi-month forklift rebuild is often avoidable, because session hosts can be retasked to a modern broker in under 15 minutes by swapping the agent service, leaving golden images, application runtimes, and local user configurations alone. The session also covered the unglamorous parts of remote delivery, legacy binary payloads and WAN latency, and how the team cleared file transmission and HTML5 redirection bottlenecks by pushing dedicated filesystem drivers to users. Passport to Protection: Entra Conditional Access Policies for Travelers We’ll admit to being Entra novices, so we were glad Daniel Keer of Digitally Accurate took on a common identity problem: securing traveling employees without administrators hand-managing temporary travel security groups. Keer, who works in Microsoft Entra ID, Intune, and Access Packages, showed how to automate travel policies without opening holes in the perimeter. Figure 7: Daniel Keer before his Conditional Access session. Standard geo-blocking policies rely on static named IP locations, which drop connectivity the moment an employee logs in from an airport lounge or an overseas hotel. The usual fix is a temporary manual bypass rule, which brings configuration drift and security exposures that linger after the trip. Keer’s answer couples Conditional Access with Entra ID Entitlement Management. Organizations publish self-service travel catalogs through Access Packages; users submit time-bound travel windows with automated manager approval and hard expiration dates. That hands-off workflow pairs with compensating Conditional Access rules, phishing-resistant MFA and Intune device-compliance checks among them, so traveling accounts stay hardened without a permanently open administrative back door. Figure 8: Entra ID Conditional Access signals: identities, locations, and device compliance. Keer also laid out identity baselines every enterprise should enforce: - Legacy auth deprecation: explicitly block legacy authentication protocols that bypass modern multi-factor controls. - Client platform hardening: restrict unmanaged and unknown client platforms to curb automated phishing-as-a-service token harvesting. - Device code flow restrictions: lock down OAuth Device Code Flow to eliminate device-code phishing and session hijacking lures. - Roadmap preparation: get ahead of Microsoft’s deprecation of legacy custom controls in favor of native external authentication methods before the May 2027 milestone. He strongly advised validating every new Conditional Access rule in Report-Only mode and with the Entra What If tool before enforcement, and exempting emergency break-glass accounts so nobody locks themselves out of the tenant. Empower Panel: Navigating Systems Engineering Burnout After lunch and a lap of the vendor floor, the afternoon opened with the Empower Panel, with Jarian Gibson, Janna White, Kai Berry-Helmlinger, and Riley Fiske. The panel took on the human side of enterprise IT: how to combat fatigue among the systems administrators who keep round-the-clock hybrid workforces running. Figure 9: Empower Panel on burnout across EUC engineering teams. Leostream: One Control Plane for Modern EUC Leostream CEO Karen Gondoly’s session, “One Control Plane for Modern EUC: Broker, Secure, Orchestrate, Anywhere,” was a technical walk through the company’s vendor-neutral approach to centralized access, brokering, and orchestration across mixed environments. Figure 10: Leostream CEO Karen Gondoly presenting “One Control Plane for Modern EUC.” During the session, Gondoly previewed Leostream Cloud, a fully hosted SaaS management plane that removes the need to deploy and maintain an on-premises connection broker and the infrastructure around it. Leostream built its 20-plus-year reputation on self-hosted Linux connection brokers that sidestep Windows Server licensing; the cloud tier moves that control plane into a managed service. Leostream hasn’t announced it publicly yet. She also walked through how Leostream uses an outbound-only agent on host targets to establish a reverse SSH tunnel back to the Leostream Gateway, which eliminates inbound firewall rules, simplifies zero-trust edge connectivity, and coordinates display protocols including Amazon DCV, HP Anyware (PCoIP), and Microsoft RDP. Protocol and hypervisor agnosticism remains Leostream’s strongest argument. It brokers physical workstations, Proxmox VE, VMware vSphere clusters, and AWS and Azure instances from one console, and its provisioning and power-management rules shut down idle compute when shifts end, trimming cloud spend while keeping GPU and CAD workstations behind policy controls. UnConference: Credit Card Points and Local LLMs The afternoon wrapped with the community UnConference, where attendees submit and upvote breakout topics on the spot. Johnny Ma led a packed room on maximizing credit card points for airline and hotel upgrades. We’re taking a journalistic mulligan on the details of that one. Figure 11: Johnny Ma’s points and miles session. The breakout we chose was Stephen Wagner’s and Jaymes Davis’s session on two approaches to hosting and running large language models locally. We’ve seen the same roadblocks whenever engineering teams evaluate AI adoption: data privacy, model governance, and unpredictable public cloud inference bills. Hyperscale enterprises can buy multi-node DGX clusters; mid-market organizations and departmental labs need a secure way to prototype retrieval-augmented generation (RAG) workflows without sending proprietary code or documents to a public cloud. Figure 12: UnConference voting. Wagner and Davis showed how local inference engines like Ollama, paired with a frontend like Open WebUI, let teams index runbooks, parse technical documentation, and build local coding assistants without the data leaving the building. We added that preflight profiling tools such as llmfit are worth running first: they benchmark host RAM and VRAM against model parameters so administrators know what fits before they try to load it. Dinner with Strangers and Milwaukee Wrap-Up Wednesday closed with Dinner with Strangers, a World of EUC event where sponsors host tables at restaurants around Milwaukee and engineers, architects, and community delegates end up seated with people they’ve never met. The day had covered DEX, Entra hardening, and broker failover; the dinner conversations drifted to local sports. We like this part of the event because it gives the people on the front line of EUC a night out with peers and vendors, something usually reserved for upper management. Final Thoughts When we heard Amplify was heading to Milwaukee, we had mixed feelings. Kevin grew up near it and always spoke of it in glowing terms, but flying in can be a logistical challenge. After several days here, its appeal is clear: riverboat cruises on the Milwaukee River, the Harley-Davidson and bobblehead museums, German and Italian dining, fried cheese curds, and Midwestern hospitality. The Cream City punches above its weight. Figure 13: Gertie the Duck, Milwaukee’s 1945 wartime mascot, on a downtown mural. After attending Amplify last year, we were glad to return as a media sponsor. Of all the industry events we cover across enterprise storage and infrastructure, conferences like Amplify, where hands-on engineering, community, and practical problem-solving come before vendor marketing, remain our favorites.