Wilco van Bragt - LinkeIn Wilco van Bragt - Twitter rssa 

E2EVC 2014 Brussels

From 30 May till the 1st of June the E2EVC Brussels event took place again. With about 150 people of 19 countries the event was again well visited and with a full agenda (including a session presented by me and Freek). I blogged my experience of the event in this blog post.

On Day 1 (30 May) the event was officially opened by Alex Juschin. The first officially session (after already 20 minutes delay) was presented by Yoni Avital of Smart-X showing their Control-Up product. I loved the way that he just starts with the product running; now PPT slides about the product. One disadvantage it's difficult to blog about it :-). He did a walkthrough how the product works; just view the video when it's available to get a good overview of the possibilities.

The second session of the day was Some Things about AppSense that you (probably) didn't know by Simon Townsend and Bas Bremmer. Simon started off with some hypes and trends are announced at the moment and how Simon sees those topics. Simon states the User Workspace simply evolved, it's never or (but both). Simon continued explaining profile management (although UEM is much more) and how AppSense is dealing with it in their personalization component. Bas demoed it by using showing application changes using a ThinApp /App-V package for the same application. Next topic was the parallel processing and just in time/case possibilities of       AppSense around policies (ADMX) and scripts. Pre-Session, Pre-Desktop and Desktop Created are three new conditions in the upcoming version. Next topic was privilege, where Simon explained the issue and Bas showed the possibilities of the product (UAC, executable blocking, disabling account a security context). Simon ended the session summarizing the 10 Things you may not have known. New versions of AppSense AM and EM will be released end of June (2014 I guess) also API's are now available.

Next session was Best Practices in VM design on vSphere by Andrea Mauro. He started with the fact that best practices are no absolutes (it's more recommended practices or guidelines). He started with VM resources, the VMware virtual hardware versions specifications and Guest OS Specifications/VMware Tools. Till know I'm missing the best practices, this is just mentioning the specifications of the components. He continued with hot-add features and tuning VM. I must see that this presentation does not fit the audience of E2EVC. Mauro also touched some performance settings like Low Latency Application Setting, RAM over-commitment, transparent page sharing, memory ballooning and memory comparison. Next topic was VMware storage, where finally one recommendation was stated on the slides, but it was followed by the VMDK specifications. Mauro ended with discusses P2V dilemmas (are there still people doing P2V).

Next session was by Michel Roth and Bob de Kousemaeker with the presentation about 5 Ways to impress your customer. They started the presentation with some customer cases where the IT stores solved the challenges (77 pages user creation document, sneakernet, helpdesk ticket reducement, savings of a lifetime by providing local admin rights, pass on password problems). After those customer cases they are showing the product in a demo. One of the attendees is picked out and created as a user in the demo company and showing the IT Store User Interface and the password reset feature. At the end of presentation they used an own created service to raffle an Apple TV.

After lunch Mike Resseler provided the session The Economics of data protection. He started with some trends in de IT (Cloud, Big Data, and Self Service). Mike describes that many companies thinks data protection is already in place. Mike started with the challenges namely back-up and what the current solution should be capable of (knowledge of VM, storage, snapshot VM without downtime; use CBT, different restore options). The layered approach of backups was discussed by Mike. Mike continued with deduplication/compression with topics like forever incremental back-up, exclusions and so on. Mike suggests: 1 solution for multiple hypervisors, 3-2-1 rule (3 copies, 2 different media, 1 offsite), automate recovery tests, item level recoveries. At the end some Veeam products are mentioned that can help you out (Mike works for Veeam).

After the session of Mike, Freek Berson and I presented our session Windows Server2012 R2 RDS versus Citrix XenApp/XenDesktop, a comparison. Hopefully the audience enjoyed the session and our Agenda as a Service (Anais) model.

Then it was time to start the break-out sessions. I had chosen the session of Gerjon Kunst about a1000 user VDI-in-a-Box. Gerjon explained the customers environment of two firms (used fictive names based on beer companies, that the E2EVC style), the reasons why the project started and the requirements (made by an IT external IT consultant). Gerjon mentioned the Pro's of VDI-in-a-Box (simple to scale, short implementation, UPM/Citrix Policies, can be used with SF/NS, DSN not required, AD not required, server sizing tool, cheap[er], simple update image). Also the cons were discussed (one vlan per image, no app publishing, license only per user, no multi domain, rebooting images cannot be scheduled, grid server needs connection with hosts and vm's, limited monitoring, internal web interface is restricted, load balancing done by Grid Appliance, image versioning is a challenge and cannot resize after image creation). Next topic were the design decisions (all local storage, except profiles, only pooled Win 7 /wo pVD, 10 GB link between sites, same desktops for both entities, additional vSphere license/API closed on free edition, Immidio for Profile Management, Zenworks for App Delivery and both datacentres active for VDI. Gerjon shows the server sizing calculator and the final infrastructure. Gerjon described the migration steps (admt, immidio for profile migration, zenworks deploy URL). Gerjon ended with best practices/lessons learned: use short DHCP lease times, do not change anything through SSL, LDAP account needs permission to add unlimited accounts of computer, refresh desktops on regular basis, sizing tool is just a guideline, delete computer accounts doesn't always work, when re-adding host to cluster re-install host, do change OU after creating image, licensing). Gerjon ended with the statement that vdi-in-a-box is scalable for large(r) environments, but the question is what's the future of VDI-in-a-Box.

Next session I attended was by Benny Tritsch with the presentation a survival guide to a Microsoft-only VDI. Benny also started with the naming convention of Micorosoft and the available clients (acquired by Itab). Next topic discusses Pooled VMs versus Private VMs. Benny shows the Management Console and talks us through (the limitations) of the console and collections. Next topic was prepping the VMs, Benny discusses and shows a nice creation PowerShell script creating the VHD(X) and own created PowerShell scripts for creating the VMs. Benny continued with HA/Scalability (1000 con/sec per RDG, 5000 conn/5min per RDC, 60 MB SQL server DB for 5000 VMs, 5 GB per pooled VM). Benny also touched the User Profile Disk, but because we already did that in our presentation he didn’t go into details. Benny also described User Experience Virtualization characteristics. Benny tries to explain MS licensing and came up with a pretty easy and very good flowchart. He also touched the licensing trick using Server OS for one user and Project VRC shows that it performs better than the Client OS. Benny ended with some good notes from the field (gen 2 VMs unsupported, no mix and match RDSH/RDVH, PowerShell is not for OPS, no integrated management console, incompatible profile management solutions, SCVMM does not cover VDI solutions, seems not to be very scalable.

Last session of day 1: Pierre Marmignon with the presentation Virtual Desktops Best Practices Chapter 3: Login Times and Response Times. Pierre explains that AD can be one of the reasons of slow logon time. He provided tips and tricks like user groups to assign resources, organize RDS/XenApp/VDI into separate OUs, leverage GPO hierarchy for common settings, if possible block inheritance at top level OU, permissions on non-standard groups are domain specific and client side extensions cross GPO impact performance. Next topic about login times were licensing (sizing, force server instead of discovery), profiles (leverage UPM, enable streaming, exclusions are not an option and need to be managed wisely) and Anti-Virus (exclude all Citrix components, do not scan the same item twice, and check Project VRC comparison). Pierre repeats that GPP should be avoided and this can be better done via logon scripts and again some tips and tricks on GPO (avoid WMI filtering) and logon scripts (forget about PowerShell, write scripts by a skilled writer, use asynchronous execution when possible, VBS is still one of the best solutions, query LDAP/AD smartly, user feedback is key, consider using third party solutions). Pierre continued with other tips for response times (OS tuning, tune Windows Explorer, enable Citrix CPU optimization).

Second day started really early on 8:30 with the session XenMobile and Beyond by Esther Barthel, Martijn Hulsman, Karen Scibberas and Rienk Spies. I run in the session a few minutes late when Esther was describing the available XenMobile editions, followed by the XenMobile architecture. After this overview they started diving into the components into detail starting with device manager. The Pre-Install checklist was discussed (some good notes about specific requirements, just use that document when installing XenMobile). Martijn provides some lessons learned: use a wildcard certificate, use your own CA/PKI for internal SSL, add root CA to NS, XDM and XAC cer store, use XDM self-signed certificates during install, request apple push notification server certificate before install, Citrix is required to sign certificate, get all DSN FQDNS in place in advance, use simple URLs, Auto Discovery. Next topic was the App Controller again starting referring to the installation documents and a description of the AppController (virtual appliances, initial network config in console, and further config via web console). Esther describes the functionality using print screens from the console and user interface. Also the MDX Toolkit (aka Worx App SDK) is described, it's only available on Mac OSX. Esther continued describing the microVPN tunnel, WorxWeb (by default not MDX   Enabled), WorxMail. Martijn discusses the Lessons Learned for the AppController (Separate URL for AppController, Direct connections to StoreFront for laptops, use an Enterprise Apple Developer Account (personal is limited to 100 devices), Android Apps also require the MDX Toolkit. Next topic was the Enterprise edition features starting with ShareFile Enterprise (Storage Zone Connectors), decisions to use CYO (Device Manager)/BYO (AppController, also includes wipes on application level). Lessons Learned according to Martijn are test each component separately before integrating, use internal PKI to enable certificate based authentication, use a specific test device. Now it's time for the demos which were provided by Rink. After a short break Karen took over talking about troubleshooting XenMobile. Karen started with describing the authentication flow, so which components communicated with each other. Karen continued with the available troubleshooting possibilities. She started with the MDM logfiles website (http://<IP>/zdm/helper.jsp). Next topics were the logfiles uploading for the Worx applications, followed by looking into that logfiles. Next topic was the NetScaler Gateway combined with the AppController and Single Sign-On. Testing via a Browser is the first step of the troubleshooting process. Last topic was the new feature in AppController 2.9 to forward users to a specific NetScaler. Karen ended the session with referrals to good documentation and takeaways.

Next session I attended was Designing & Building Hyper-V over SM3.0 by Aidan Finn. Aidan started explaining where the session is about and which components/hardware is involved (Storage spaces, 2-8 clustered file servers, Cluster Shared Volumes, File Server for Applications Data, File Shares, and SMB 3.0). The hardware (JBOD) is by less known manufacturers normally. Aidan discussed the hardware components for this solution (its physical solution because of SAS connectors). Next the Storage Pools and the disk characteristics were described. Next Adain does a walk-through of the installation and configuration steps he performs (check the video or slides for all details) both based on slides as recorded demo's.

Just before lunch Mike Resseler was on stage with the session Data Deduplication in Windows Server 2012 R2. Mike started of describing Data Deduplication and situations were data deduplication is already in place without noticing like the SSX folder. Mike explains the differences possibilities of data deduplication: in-line v post process, source v target, file- block- or chunk based. Next topic was planning with topics like determine target deployments (works for file share, virtualization files, os deploy files, back-up files, could be for LOB, web servers, HPC), hardware requirements (, unsupported configuration, evaluate savings and policies. Mike mentions the minimum requirements of the machine, followed by the configuration steps. Mike mentions the default settings (3 days last modified, schedule every hour, exclusions edb/jrs, garbage collection/data scrubbing on Saturday morning. Mike continued with the architecture (filter drivers, service based [ddpsvc), garbage collection, optimization. Mike mentions the jobs behind quickly, the settings are summarized again and where you should think about. Because it's using caching by deduplication better performance is reached. Last topic is using Deduplication for VDI. The session ended with a summarization with the (MS mentioned) benefits, performance hit (caching will leverage during time), reliable, can be used with branch cache, works with DFS Replication) and the evaluation tools.

After lunch it was time for the Andrew and Jim show called Driving Crazy Fast VDI. As the session is about Atlantis so they started with their product range (ILIO Persistent VDI, Diskless VDI, XenApp, USX), the characteristics (software only VM, any hypervisor, any server/storage platform, server-side inline deduplication) and how it works (architecture). You can test drive the product via the online demo environment. Jim touches the results of VMware and MS which are achieved by 100% read benchmarks. Next topic was the free ILIO appliance for 10 clients, which was shown in a (recorded) video. Andrew continued with explaining USX, followed by the upcoming USX Enabled Data Mobility.

Next session I attended the presentation of Ingmar Verheij and Kees Baggerman Podio, the next gem of Citrix. They started with some facts about the acquisition together with questions for free t-shirts. Next they showed the video of Citrix Podio and some quotes of the product. Podio is a bit of Facebook, yammer, dropbox, sharefile combined together. I'm mentioned a bit too much in this presentation that I can work with it :-). Next Kees explained how Podio is structured. Ingmar showed Podio in real life creating an application and the apps we are using within the DuCUG. Kees summarized what Podio can do using Will it blend. Ingmar explains the license model and what the free version is offering. The session ended with some good Q&A with the audience.

The last session I attended of this E2EVC Brussels was the session of Helge Klein How to shine as a consultant with the free edition of UberAgent for Splunk. Helge started explaining what Splunk is and how it works. Next Helge explains what his product UberAgent does and how it cooperates with Splunk. Helge would like to show his reports, but it did not work directly. But he got many good questions about Splunk and UberAgent, maybe this is arranged by the beers and chocolates he was giving away for good questions. Helge fixed the issue and shows some of his Dashboards and explains the data shown on those dashboards including some enhancements in an upcoming enhancement.

Because of conflicting appointments I could not attend the third day this time. Luckily most of the session’s presentations on Sunday this time, I already attended on earlier events, so I just missed a few sessions (sessions are recorded so I can view those later on). Again the E2EVC team did an amazing job by organizing the event. It was again filled with good sessions and as always there was a great group of people who attended the event. I have met again with new people and old friends and enjoyed the conversations a lot. Already looking forward to join a next E2EVC event. The autumn edition of 2014 is held in Barcelona, for 2015 again Brussels and Rome are scheduled.