In the attached link you can find a list of the major improvements in vSphere 5.5 summarised by VMware.
Everyone will have features that will be most relevant, or favourite, either due to personal experience or active projects being worked on. When you look at what’s on the list, there are a lot of very significant substantial improvements.
The doubling of most configuration maximums is significant and finally one of the most important improvements is the increased VMDK size of up to 62TB. Note that any virtual machine specific changes require VM v10 hardware. See caveat below.
Maximum VMFS-5 size is still 64TB but the reason for a maximum VMDK size of 62TB is for storage of VMDK metadata, snapshots and redo logs, so as not to fill up the datastore. For larger database systems like SQL Server, Oracle and SAP this is significant as 2TB+ VMs are becoming more and more common.
We all need to understand some of the limitations of the Web client and traditional VI (c#) client. At present until the next release of vSphere when the VI client will no longer be available, there are some interoperability issues.
Examples are the ability to only manage VM V10 hardware using the Web Client – it is not supported on the VI Client. Despite this there are still limitations with what is supported on the web client, such as only limited support for Update Manager. You cannot create Baselines and Baseline Groups for example, using the Web Client. When the next release of vSphere is available I think we can finally see a lot of these issues resolved.
Anyone who struggled a bit with SSO in vCenter 5.1 can expect to see big improvements due to a much simplified architecture.
VSAN is also a very significant enhancement and provides a very realistic VMware distributed storage model. One of the nice things is that overhead on a host when this is implemented so far is expected to be maximum 10%, but in reality much less than that.
Finally, the previous handling of Storage “All Paths Down” or APD conditions has been significantly improved. In vSphere 4 this sometimes caused disconnects from vCenter to ESXi hosts. Now connectivity with a storage device occurs, it enters a “Permanent Device Loss” or PDL state where it will be removed from any connected hosts.
There are so many improvements so best thing is to get it in the lab and test it out.