One conversation I find myself having a lot with customers right now is the “LTSR vs. CR debate.” Anytime we ship a new Citrix Virtual Apps and Desktops CR with another “killer feature,” the debate generally heats up. And when the CRs start to stack up before the next LTSR, the debate really heats up.
I think there are three reasons why customers are asking me (and other Citrix Consulting Services colleagues) for advice about this right now more than ever:
- We’ve shipped five CRs since the last LTSR: 7.16, 7.17, 7.18, 1808, 1811.
- The next LTSR is still maybe a year away. (We’ve announced our intentions for H2’19 but that could still mean a few more CRs.)
- The killer features between the five CRs are really starting to add up (and we all like new, shiny toys).
Before I get to my top eight killer features that might make you consider adopting the CR track, I want to emphasize that your choice of LTSR or CR should really be a business decision. It’s not about whether an admin or super user wants feature X. It should come down to where the business ranks supportability in terms of importance, whether the business requires new platform support for a mission-critical workload (i.e. Server 2019), and whether IT is agile enough to keep up with the CR track, risk tolerance, and more.
Let’s face it. We all want the new, cool features from these releases. But there is a big difference between wanting something and truly needing something. And it’s not like this decision or debate is something new. We had seven CRs after the 7.6 LTSR, and there were some amazing features like zones, app groups, and LHC. I always advise customers to lay out the pros and cons of each approach and let the business requirements drive the decision.
I admit that I sometimes get “labeled” as a big LTSR advocate. In my defense, that’s probably because I spend 99 percent of my time working with large, enterprise customers who typically value supportability and uptime more than anything else. And I think LTSR makes a lot of sense in that case.
But if you’ve mastered the world of automation or the business decides feature X is a hard requirement, then the CR track might be for you. Go for it! I wish more customers could confidently adopt the CRs we’re putting out because the killer features are piling up.
So here are my Top 8 post-LTSR items that might make you think about adopting the CR track:
1) Key UX Enhancements
EDT is a game-changer, as I’ve talked about in the past. There, we essentially flipped ICA to run over UDP vs. TCP by default in 7.16 and set the stage for the future. We took a page out of Framehawk’s book and decided that bandwidth is ubiquitous today. We’re now trying to intentionally saturate the pipe instead of being “thin” on the “wire” (check out our original Thinwire patents for some good bedtime reading). That was a smart move and laid the foundation for adaptive throughput in 1808 and 1811.
When I started at Citrix, offshoring was very popular, and we’d battle latency in the 300-500 ms range all the time. One of the tweaks we used to make was changing the number of output buffers that ICA/Thinwire used. We’d generally bump it up to a magical value of 118, and that made things better in WAN scenarios. Check out this whitepaper we published in 2003. Page 11 shows we recommended tweaking the “OutBufCountClient” parameter to a value of “118” in the ICA template file. Fast-forward 15 years, and in the 1808 release, we finally changed that value permanently to 100 (default = 44 in case you were wondering).
But you might be asking if that is optimal in all network scenarios or on a LAN. The answer is no. So in the 1811 release, we put in some much-needed smarts to dynamically adjust this value based on network conditions and user interactivity. The other thing we did that didn’t get a lot of press was we moved this config from the client to the VDA (easier for us to control and no more tweaking client-side regkeys or parameters). But you add it all up and our protocol is getting smarter and smarter, and the user experience is getting better and better.
And, finally, another feature that customers couldn’t live without a decade ago is back: Local Text Echo! This didn’t make headlines in the 1811 release, but I know all the old-school Citrix administrators out there will know and appreciate its return. It was first introduced in the MetaFrame XP days as SpeedScreen Latency Reduction (SLR). Along with instant “mouse click feedback,” Local Text Echo would kick in if your latency reached 500 ms or higher (but we’d often reduce that value to 150-200 ms). The idea was pretty simple — when you type something in a session (i.e. composing an email in published Outlook), it renders it immediately on the endpoint, with no lag or perceived latency. And we process it server-side in the background. It was instant gratification and users loved it! Brian Madden loved SLR so much he dedicated a lengthy piece to explaining SLR and how Local Text Echo worked. Long story short, Local Text Echo disappeared several years ago as we made the transition to the FMA architecture. But now it’s back and better than ever.
2) HDX Insight 2.0
This is another game-changer. In 7.16 (and NS 126.96.36.199), we completely re-architected how HDX Insight works. We introduced our 28th virtual channel, called “NSAP,” which provides a dedicated, uncompressed path for Insight data. This can easily double the scalability of Insight/AppFlow, not to mention improve the reliability and stability of your NetScaler/ADC boxes. Read more about it in my HDX Insight 2.0 post.
3) Outlook Search Redirection
This is easily my favorite feature from 7.18. We finally solved how to deploy Outlook in a non-persistent virtualization world — with native search capabilities! We did it with Profile Management’s large-file handling feature and by taking a similar approach a lot of the other vendors out there took (FSLogix, etc.). We use VHDX files/containers to redirect the Outlook OST and search index database files, along with the user profile. That way, we don’t have to rebuild indexes, logons are fast, and these critical files persist between sessions. We finally have a really good answer for “O365 + Citrix” that doesn’t involve a third-party solution.
4) PVS Asynchronous I/O
This is something I’ve been talking about with the PVS engineers for three years. Until PVS 1808 debuted, a target device served incoming OS storage requests by traversing through three different layers (RAM cache, VHDX file and network streaming) sequentially to complete a request. With 1808 (and a checkbox to make your life easier in 1811), we can now send and process the three streams asynchronously, which reduces latency and greatly improves performance. It didn’t get a lot of headlines, but if speed is king and you’re using PVS, I’d encourage you to take this hidden gem for a spin.
5) User Layers
We acquired Unidesk and announced deprecation of Personal vDisk and AppDisks in 7.13. But we didn’t really have an answer when it came to “hybrid VDI” with user-writable layers. Well, with the release of App Layering 4.14 (right after 1808), we promoted User Layers from Labs to GA. This is a huge feature because it replaces an entire product long-term in AppDisks/PvD. And with every release since 4.14, we’ve made User Layers better so it might be time to take App Layering for a spin (or finally replace your unsupported PvD or AppDisks environment). Just note that User Layers mount at user logon as opposed to machine boot (as PvD and AppDisks did), so this will require some level of testing.
6) Thin Provisioning of Block Storage
Seven years ago I wrote a post explaining why XenServer kept thick-provisioning block-based storage (even if the underlying storage supported thin provisioning). It really came down to our underlying storage mechanics in XS. But with XenServer 7.6 (1808 timeframe), we finally added official support for thin provisioning of block storage like FC and iSCSI. Thank you GSF2!
7) Director Drill-Downs
Logon times are very near and dear to my heart, and I get asked what a “good” logon time is all the time (and I usually cite the “Rule of 30” I made up almost three years ago). Well, between 7.18, 1808 and 1811, we added three additional drill-down capabilities of Director to further dissect logon times. We can now see what’s taking so darn long within Interactive Session; we have visibility into GPOs; and we can even inspect the profile load portion (profile size, number of large files, etc.). I’d probably file these under “nice to have” instead of “must have.” But fast logon times are still critical, so these are welcome additions.
8) Server 2019 Support
Last but not least, with the 1808 release, we added support for the Microsoft Windows Server 2019 platform. As I mentioned earlier, if the business is demanding virtualization for a new, mission-critical workload that might only be able to run on 2019, then your hands might be tied and the CR track is the one for you … whether you like it or not!
Of course, there have been a lot more features than these eight in the last five CRs. But these are some of the things I’ve been talking about most frequently with customers who are running the 7.6 or 7.15 LTSRs and who inquire about adopting the latest CR.
I might recommend using this as a checklist when having a healthy LTSR vs. CR debate. You can quickly determine which features are irrelevant, which would be nice to have, or which you can’t live without. That can feed into your ultimate list of pros and cons that you can present to the business so they can make an informed decision.
Again, I hope your list or report includes things like supportability, user experience, automation expertise, risk tolerance, and more. After you’ve evaluated all those with your unique requirements in mind, I believe you’ll be able to make a sound business decision.
I hope your new year is off to a good start, and I’m looking forward to seeing many of you at Synergy in May!
Nick Rintalan, Principal Architect – Citrix Consulting Services (CCS)
Citrix Tech Bytes – Created by Citrix Experts, made for Citrix Technologists! Learn from passionate Citrix Experts and gain technical insights into the latest Citrix Technologies.
Want specific Tech Bytes? Let us know! firstname.lastname@example.org.