In my last post, we discussed how the three EqualLogic hybrid iSCSI SAN arrays are positioned for different desktop virtualization use cases. In this post, we focus on VDI with EqualLogic hybrid blade array.
Dell EqualLogic PS-M4110XS arrays are half-height, double-wide hybrid blade arrays with five 400GB SSDs and nine 600GB 10K SAS drives for a total of 7.4 TB (raw) capacity. Even with its compact, modular design to fit within the blade chassis, the PS-M4110XS blade functions as a full EqualLogic iSCSI SAN array, with all the capabilities EqualLogic has become known for.
These hybrid arrays provide a blade form factor suitable for a comprehensive, self-contained VDI solution within a modular and compact blade enclosure. Together with Dell PowerEdge™ blade servers and Dell Force10™ blade switches, EqualLogic PS-M4110XS hybrid blade arrays create “converged” data-center-in-a-box VDI deployments.
What does it mean? Why does it matter?
For a thorough answer to these questions, check out this recent paper from the Taneja Group.
In a nutshell:
- The entire VDI solution – compute, networking and storage – delivered within the blade chassis. No external switching for storage, very easy to manage, a whole lot less cabling – an extremely space-efficient, dense, cost-effective VDI solution.
- Unlike other “converged” solutions in the market (which are essentially a pre-tested racking and stacking of the infrastructure components), you can achieve a true “converged” solution where you can reliably run SAN and LAN traffic together through the same switch with a QoS guarantee. DCB-enabled Dell Force10 MXL switches and EqualLogic PS-M4110XS hybrid blade arrays make this a very attractive converged VDI solution that has a very low dollars-per-desktop cost.
The other question we get pretty much in every customer conversation involving VDI with PS-M4110XS arrays is this: “What is the optimal VDI configuration for the solution – blade servers, blade switching and blade arrays together – and not just for the hybrid blade storage.”
The Taneja Group paper mentioned above has some sample solution architectures for this VDI solution.
In one use case, we can have two hybrid PS-M4110XS arrays for a total raw capacity of 14.8 TB per array group. As shown in Figure 1, that leaves room in the chassis for 12 x M620, two-socket, half-height server blades (see front panel view at top left of diagram), with 2 of those M620s dedicated to infrastructure and management. The middle and right columns show back panel views, including fans, power supplies, and the vertically oriented Fabric IO Modules. In the middle column scenario, DCB is enabled to converge LAN and SAN traffic; and in the right column, the LAN and SAN each run on a dedicated fabric.
Figure 1: Converged VDI solution with Dell blade servers, blade switches and blade storage
A converged infrastructure VDI solution with two hybrid PS-M4110XS arrays can support 800 to 1000 typical task worker desktops. (Of course, the exact mileage will vary depending on the VDI workload your environment is creating.) In this solution with two blade arrays and 12 x M620 blade servers (with 2 allotted for infrastructure and management), each M620 blade server can be used to support around 80 to 100 virtual desktops, depending on the application mix.
How about the performance of this converged VDI solution?
We have tested this VDI solution with PS-M4110XS hybrid blade arrays and some of the results are included in the Taneja Group paper. Figure 2 below shows the boot storm results for 450 desktops running on a single PS-M4110XS hybrid array (the full solution will have two of these).
Figure 2: Boot storm test: 450 desktops on one PS-M4110XS array
For full details, check out the Taneja Group paper.