We are a hospital replacing 1500 desktop PCs before our new campus opens in the fall with thin terminals connecting back via RDP to 1500 WinXP VMs as a VDI solution.

2 datacenters

6 HP c-class blade chassis, 3 in each datacenter.

38 HP ProLiant BL480c blades with dual Quad-Core Intel® Xeon® 5300 Series processors and 24GB memory (3GB per core)

8 blades per chassis

4 NICs on each blade, 2 to each internal chassis cisco switch (one uplink each to core datacenter cisco 6513).

VLAN VMotion and Service Consoles (seperately)

2 EVA8000, 1 in each datacenter. 2 clusters will use "EVA1" as primary storage with CA to "EVA2". 2 clusters will use "EVA2" as primary storage with CA to "EVA1".

38x300GB FC drives (one diskgroup), 22x500GB FATA drives (another diskgroup)

4 ESX clusters with DRS & HA

VDI-A: 10 ESX (5 in each datacenter), 400 VMs (40 VM per ESX with room for complete datacenter failover: 80 VMs/ESX with 5 ESX)

VDI-B: 10 ESX (5 in each datacenter), 400 VMs (40 VM per ESX with room for complete datacenter failover: 80 VMs/ESX with 5 ESX)

VDI-C: 10 ESX (5 in each datacenter), 400 VMs (40 VM per ESX with room for complete datacenter failover: 80 VMs/ESX with 5 ESX)

VDI-D: 8 ESX (4 in each datacenter), 320 VMs (40 VM per ESX with room for complete datacenter failover: 80 VMs/ESX with 4 ESX)

8 datastores per cluster/ESX host (one LUN per datastore)

512GB LUNs (50 VMs with 10GB C: drives + config/log/vm swap \[50 * (.5*512MB)=12.8GB swap])

VMs are WinXP SP2 with 512MB memory, 10 GB C: drives. Altiris used for imaging & customization (initially deployed from templates w/customization spec). Appstream used to stream apps.

Did I miss anything? Can anything be improved? We are looking for failover capability as well as management efficiency.