If you would like to read the other parts in this article series please go to:
Introduction
Citrix recently released the first Citrix-branded version of their VDI-in-a-Box product, which is the next version of the product the company acquired with Kaviza. I was fortunate enough to be able to spend some time interviewing Krishna Subramanian, Vice President, Marketing& Business Development for Citrix Systems, Inc.
I’d like to thank Krishna for her time and for her excellent answers.
This is the second of this two part interview.
The interview – continued
Scott: Do you see VDI-in-a-Box competing with VMware View or Microsoft, or are those generally just more enterprise level? Or would those compete more with XenDesktop, in your opinion?
Krishna: VMware View is one that our customers have tried mainly because if they have virtualized their servers with VMware, they already have familiarity with the brand, so they always try something like View. I think the challenge for View in this market is that it is an enterprise solution. The challenge with VDI is 60% of the cost of VDI is not in the software, it is in the physical infrastructure. Even if someone were to give you VMware View for free, it still would be much more expensive to deploy that VDI-in-a-Box because of the physical infrastructure that it requires. When we see something like View in an account, we typically end up winning those accounts because of this cost advantage. A lot of customers who are looking at our solution have probably looked at something like View and have looked at things like terminal services, but need something more than terminal services.
Scott: Do you have linked clone functionality?One of the big concerns about VDI is that it can eat up a ton of disk space.
Krishna: We do linked-clones by default.
Scott: Is there a preferred hypervisor for VDI-in-a-Box? Do you care, or is there a preferred hypervisor that it works best on?
Krishna: We don’t care. We’re hypervisor agnostic and with the version 5 release, we’ve added support from Microsoft Hyper-V, then server, then VXE and VXI.
Scott: Could you scale per server a little bit more the hypervisor features that are built in like VMware’s memory sharing techniques? RAM is generally the most constrained resource in a virtual environment except for a VDI environment where you have a bit more performance. Do some of the hypervisor features help scale VDI-in-a-Box or does that not really play into the equation?
Krishna: We do leverage all the features of the hypervisor, so, for example, with the XenServer, even when you use the free version of XenServer with VDI-in-a-Box, we upgraded, at no extra charge, to the paid version of XenServer, and you get all the features of the paid version. You don’t get the support for the version, but you get all the memory, the optimization of XenServer when used withVDI-in-a-Box. So we do leverage all those capabilities of the hypervisors when we do the virtual desktops.Then we have our own optimizations becauseVDI-in-a-Box has a pretty unique architecture because we don’t require shared storage. We have a way by which we optimize use across the different servers withVDI-in-a-Box grid, and we do use our own implementation of linked-clones as well to save on storage.
Scott: In order to provide availability and redundancy, you’ve replicated across the grid?
Krishna: Yes. We replicate intelligently across the grid.
Scott: Can you tell me what you mean by “intelligently”?
Krishna: If there are multiple servers in the grid, then as the grid gets larger, we will not replicate everything on every server. We only replicate as required both for high availability and for serving up the number of desktops. Depending on how many desktops you need from a template, it will accordingly replicate.
Scott: So basically, users are not connecting to a server, they’re connecting to the grid.
Krishna: Logically, they are connecting to the grid. The interesting thing about this architecture is, every server in the grid acts as a master, so there is no master-slave in our architecture. You can actually connect to any server in the grid, and that server will know whether it needs to serve up your virtual desktop or if another server in the grid should the one serving your virtual desktop and will reroute your request accordingly. So every server in the grid is a connection broker. Every server in the grid does load balancing, and every server in the grid is a desktop server.
Scott: That’s basically a built-in brokering type functionality?
Krishna: That’s correct. You can think of this as a distributing grid that acts as one logical entity. It is truly a distributing architecture.
Scott: You’re using the Citrix Receiver client?
Krishna: That’s correct. We work through HDX. Any device that runs to receiver, you can use to connect to VDI-in-a-Box.
Scott: That would include a web browser? If somebody was at an internet kiosk somewhere, could they connect through the web and get their desktop?
Krishna: Yes, absolutely. We also support RDP, so if you don’t have a receiver at an endpoint… Let’s say you’re at an airport and you don’t have HDX on that device to a browser, you can still connect to VDI-in-a-Boxand it will default to RDP.
Scott: Does VDI-in-a-Box or HDX support the use of isochronous time critical USB devices like webcams using Skype or things like that?
Krishna: Yes. Basically we support all features of HDX. HDX includes USB redirection, multimedia redirection. That’s something called adaptive orchestration where depending on your endpoint, connectible latency, and the capacity of your server, it will decide where to route or what. All those features are included because we’re essentially using the full HDX stack with VDI-in-a-Box.
Scott: HDX is, in a number of ways, better than PCoIP, and, from what I understand, better on both WAN and LAN.
Krishna: Yes, that’s correct.
Scott: How do you license the product?
Krishna: It’s licensed for concurrent users. There’s a perpetual license price which means you pay the cost one time, and you own the license outright.
Scott: There’s no difference from a licensing prospective for a pool of desktops versus an assigned desktop?
Krishna: That’s correct. There’s no difference between a pool or persistent desktop. You just pay for the number of concurrent users you plan to have on the system.
Scott: Do you have different VDI-in-a-Box editions, or is there just one?
Krishna: There’s just one edition.
Scott: Is there anything else I should know about VDI-in-a-Box?
Krishna: If you, or your readers, are interested, you can download the free trial of VDI-in-a-Box 5. You just go to Citrix.com and on ‘products’ click on VDI-in-a-Box and download it and you can try it out.
Scott: Krishna – thank you very much for your time!
Summary
Personally, I’m intrigued by what VDI-in-a-Box brings to the desktop virtualization equation. By targeting a specific market segment and need, Citrix is helping to make VDI accessible in a couple of different ways:
- A single and affordable price point reduces overall solution complexity.
- Straightforward and simple technical deployment that makes the solution accessible even for small environments.
- High availability features baked into the product that makes ongoing solution operations easier.
The first two parts of this multipart series focused on my interview with Krishna Subramanian. In part 3, you will learn about what factors you need to consider before you move on to installation, which will be covered in Part 4.
If you would like to read the other parts in this article series please go to:
