Already the last day of Citrix Synergy 2018. Just as the other days day three is started with a general session called supersession. This one is provided by Michael Lewis. After the supersession a whole day of breakout sessions. In this last day blog my experiences.
Michael Lewis was together on stage with Tim Manahan. They started off with a kind of interview style where Michael started explaining how his career started off and how it went further. They continued with the topic about trading Microsoft stocks by setting up a quicker network (a few milliseconds) between exchange trades next to the original network gaining advantage in het market. Based on his books written they talked through the new topic about baseballs teams and the different budgets which did not make a difference in the score results and writing about a team using analytics for selecting the players in stead of the “old way”. They jumped to his latest book about basketball where analytics should know much more information and information could not fit in the model of human elements and that kind of elements were added to the analytics. They continued with adding analytics within other businesses. Lewis mentioning why this is happening and several examples about using analytics.
First session after the super session was Multi-Site Scenarios Best Recommended Practices by Jarian Gibson, Shane Kleinert and Dave Brett. Dave started of explaining what a multi-site actually is (spreading service across multiple servers, active/active, active/passive, active/active [pinned]). Shane continued that besides Citrix you should think about a lot of technologies that need to be setup in a multi-site configuration. Dave continued explaining the multisite user flow for all three scenarios. Dave walked through a nice flow diagram to determine your multi-site design. Shane continued with multi-site and Citrix XenApp/XenDesktop zone considerations again with a decision tree diagram and some examples how it could look like for the possible scenarios (single site – multi zone, multi-site single primary zones, multi-site – multi zones, single DC – multi-site -primary zone). Jarian continued about the SQL options (basic availability group, always on access group, local host cache). StoreFront options and challenges were described by Dave including synchronization of app subscriptions and HDX routing. Dave continued with Traffic Management options, followed by User Persona explained by Jarian. Next topic was Image management (non-persistent MCS/PVS, persistent Full Clones, App Layering, Veeam and Nutanix were mentioned as solutions). Jarian continued with ShareFile and XenMobile. Dave took over again what happens when a failure situation occurs dependent on the multi-site.
Next session I attended was Citrix App Layering: Top 10 Lessons Learned of Rob Zylowski and Dan Morgan. Session started about the ELMs high availability and ELMs for packaging applications. Next topic was the OS layer (start with clean install and install only the required supporting components [Windows Updates, .Net, Visual C++, Flash, Silverlight, optimizations], use one OS layer per OS type, store a copy of a pristine OS layer prior to modification, store apps supported as of 4.1), followed by Windows Updates (Windows Updates, WSUS or offline standalone update packages). About the platform layer Dan discussed the priority of the platform layer, components in this layer are GPU drivers, RDS, Domain Join, PVS Target Device, VDA, Receiver, WEM, WMI scheme change, anything impacting the logon stack and modifications in the user profile. Next topic was Antivirus (use recipes, put in App layer, scan completed images not layers, consider delay before user logon, on access scanning only on XenApp/XenDesktop, consider host-based antivirus solutions. Rob took over and continued with Office (follow the recipe, only via a layered image, Visio/Project handling, O365 user Layers Stores .OST file and streaming files). PVS/MSC were the next topic (use your gold image to create a template, give machines a few minutes after boot (SettlementPeriodBeforeUse setting, MCS disable handling Windows and Office activation, PVS use the XenDesktop Setup Wizard). Rob discusses elastic file share considerations: HA strategy (Windows Clustering, SMS3 share from NAS, don’t use DFS for share namespace), scaling the elastic layer share (IOPS requirements will depend, use POD for large environments, deploy in controlled batches), Elastic Layer share path can be changed by registry on VDA, file share performance is important, layers mount once per VM reboot cycle. Also tips and tricks were shared for speeding up packaging (use the connector cache, relative difference between connector speeds, resource bottleneck can be storage speed, network bandwidth/latency and CPU speed/contention).
The last session I could attend before flying back was 4 Real-world lessons learned from my ShareFile deployment. The session started off about why using ShareFile (increase productivity, collaborate, modernize/organize the IT infrastructure, flexible integrations and protecting data). ShareFile can be implemented on Cloud, On-Prem or Hybrid. The continued what end-users expect (a plan, understanding how/when to use ShareFile, clear communication, automatic access and so on). Next topic was the 4 (actual 5) lessons learned in deploying ShareFile. Fraser Robertson of BCiT started explaining why the have chosen ShareFile (modern file storage/sharing solution, secure external file sharing, anytime/anywhere access, fitting in the existing environment and data storage flexibility supports compliance). He continued with their initial plan (engage with Citrix technical implementation to ensure they did it right, pilot groups and communicate via their communication platform. He continued with their lessons learned: 1) Testing, testing, testing (and check all use cases/applications), 2) partner with a trusted advisor with knowledge to speed and now the set-up is right, 3) consider the end-user experience (understand how your users access the files/ sync client and/or drive mapper), 4) communication is king, 5) Data Migration (understand your options [data migration tools, use individual tools, API/scripting], choose the path for your organization). Next time they will extend the pilot with all uses cases and communicate trough all available channels. There is a customizable communication kit available (link on the slides).