E2EVC 2015 Lisbon Day 2
- Published: Sunday, 15 November 2015
On the 14th November 2015 the second day of E2EVC Lisbon took place. It’s bit like the show must go on after the tragedy in Paris last night. My day 2 experiences can be found in this blog.
The day started with a session by Ingmar Verheij about the Internet of Things which about connecting everything with everything. Ingmar started explaining why he started with IoT, getting insights in the power consumption of his house and the regulations of the heat reduction and humidity in the house. Ingmar continued with the starting points: things (Arduino, raspberry Pi, intel Edison), GPIO, electronics. With some cool nice demos Ingmar showed the possibilities of IoT.
Second session of the day I attended was XenServer deep technical dives by Robert Breker. Robert started with the percentages of hypervisors used for XenDesktop 7.x and the popularity of XenServer on Cloudstack and Openstack. Robert also touched that the NetScaler MPX is running on top of XenServer as well. Robert continued with the releases of this year XenServer 6.5 with SP1 (64bit, network trroughput, quicker network establishements, storage throughput, more running VMs, memory based read caching for NFS). Robert continued explaining containers (Dockers) and how XenServer handles those, followed by GPU pass-through for Linux guests, XenServer supports Nvidia, Intel and AMD. Next topic was XenServer Dundee Technlogy Preview: SMB virtual disk storage, open-FCoE, NFSv4, offline virtual machine migration, Intel GVT-G, citrix insight services integration automated, improved hotfix application process, new hardware support, improved dom0 responsiveness, improved active directory scalability). The session ended with some Q&A.
In next set of breakout session Bas and I presented our session XenDesktop FailOver (im)possibilities. We had a full room with people with people sitting on the floor. Hopefully the audience enjoyed the session.
After our session Alex announced the events of 2016 including the special prices in Singapore, Las Vegas and Dublin.
After lunch I attended the session of Roy Textor HDX3D Pro with SuperMicro/Nvidia Bundle. Roy started mentioning the specifications of the GRID K2 and the Microserver. Next Roy did a walktrough of the XenServer installation including the GRID package, followed by the GPU configuration via XenCenter (assing vGPU profile per VM. You need to install the driver of Nvidia in the VM (graphic mode and mousecursus change). Roy continued with some optimaizations (use at your own risk): BIOS -> disable c-states/disable 64bit MMIO, Maximum Performance (differs from BIOS to BIOS), XenServer -> cpufreq=xen:performance (reboot required), vGPU: vSync, Disable Frame Rate Limiter, VM (4 vCPUs Windows Client OS support a maximum CPU sockets, you need cores, roys advices to use 1 socket met more cores as it looks that performs better), ThinClients: should be HDX 3D Pro certified, should support deep compression decodin, multimonitor support.
Next up were Barry Schiffer and Bram Wolfs with the session Citrix display bells and codec whistles. They started with showing the different graphics possibilities how they are shown in the Citrix policies. Barry mentioned and described the HDX display modes: Legacy Graphics mode (XenApp 6.5 adaptive display comparable XA65, Win7/2008R20), Desktop Composition Redirection (windows client OS VDA only, Windows OS Receiver only +GPU), Framehawk (NS11, requires UDP, 7.6 FP3, windows and IOS receivers only, requires virtual channel), H.264 (recent Receiver version, requires local video decoding). ThinWire Plus (any receiver version, JPG compression and bitmap caching technology). Bram created a utility called HDX Display Analyzer showing the current display mode with important information. It’s also possible to directly select another display settings, which are applied directly. Incredible tool again by Bram. They showed the tool with some demo’s showing differences scenarios and see big differences in CPU and bandwidth usage. Summarized you need to choose the codecs/display modes carefully for each specific situation.
Benny Tritsch and Andreas Fleischmann were presenting the next session System Monitoring with LakeSide SysTrack and App Containerization with FSXLogix Apps 2.0. Andreas started off with quickly explaining FSLogix Apps 2.0 components and basic concepts via good demos. Powerful rules for different user environment configuration and lockdown, but you need to know that to hide/redirect/faking for each specific applications. Benny took over with a technical introduction to Lakeside SysTrack. Benny started with the architecture overview of the product including the different dashboards. Benny showed the product in a demo including the available profiles.
Next session was about Citrix Workspace Cloud presented by Gerjon Kunst and Mick Glover. Mick started explaining the Citrix Workspace Cloud and which services are there (Apps & Desktops, Mobile Mnaagmeent, Secure Documents and Lifecycle Management). Via print screen Mick did a walk-through to the Workspace Cloud (web)consoles, followed by the Workspace Cloud Connector, which connects the local VDA’s with the Citrix Workspace Cloud. Mick touched quickly the Browser Service and AppDNA express beta/alpha. Mike also mentioned the options for Cloud-Hosted vs. on-Prm StoreFront including their characteristics. Gerjon took over with Citrix Lifecycle Management by explaining the possibilities for installing Citrix environment along best practices with the option to add your own scripts. However other products can be done as well, several companies already added their blueprint. CLM exists of Management Portal, CLM engine, agents, designs (installation scripts and files) and templates. Gerjon showed the steps of setting-up the CLM using printscreens. Gerjon also touched the licenses which are not fully clear at this moment and ended the session with a nice summarization of the (im)possibilities with Citrix Workspace Cloud.
Last session of the day was an extra session that Barry and Remko just prepped one hour in advance. The session started with a great introduction. The quickly switched to a demo with K2 in pass-trough mode where Barry is showing his case that browsers were still using CPU while a GPU is available. Barry showed how Firefox could be reconfigured to use the GPU better. Remko took over explaining how they searched if Chrome is using the GRID card. They ended up this was RDS related showing the display adapter different and actually is using the GRID (they mentioned a required policy for W2012R2 to support more options than W2008R2). Conclusion is that Chrome don’t work as good as IE as Firefox with GPU. Compliments for Barry and Remko providing such an interesting session in just a short time preperation frame.