Need VMS Server Setup Help - Recommended Specs And Configuration

Hi I am looking for answers for the below

1:The role of most VMS is to just send the stream received from the cameras to the storage rite.So if we take the same cameras and send the same main stream to multiple VMS like Milestone,Axxon,Genetec,Avigilon etc via ONVIF the exported footage without alteration should give the exact same quality rite?

2:CPU,GPU,Memory usage.No matter what CPU or GPU you use, as long as the usage is not exceeding 70-80%, the footage should be seemless rite?

3:The stream connection recommded is over TCP or UDP?

4:What settings of shutter/iframe/fps is recommded for clear image when pausing/frame freeze a video of a moving person to get a clear image of the face.

5:If disk configured as JBOD or RAID,fragmentation can occur? Raid 5 or Raid 6 Recommded?

By exported footage, do you mean the video that is initially stored by the VMS? IF so, then yes, each VMS will store the footage from any same camera and its quality will be the same. We see this on our own setup having different cameras simultaneously recorded by Avigilon, Exacq, Genetec, etc

If you are talking about exporting the footage outside the VMS, there can be variance depending on what CODECs and how lossless the exporting compression is set. Each VMS will have a different manner of lossless exporting.
Yes, each CPU or GPU and VMS is not created equal. So certain combinations will get you further along as far as the amount of throughput processed without loss. However, all other things being equal, any VMS that is running on hardware that is not overloaded, will record seamlessly.
This will depend per VMS setting and the network. If the video is having to stream through a firewall connection TCP is going to be required. UDP is going to create a lower stream bandwidth for a VMS to process, however technically it is not a lossless communication protocol. The problem with TCP in a high loss network is that video frames can reach the VMS out of order, and when viewing a person it will actually look like they are jumping around on the image as the frames were retransmitted by the camera if they didn't reach the VMS in order. There is no one protocol fits all answer here.
What is the rate of movement for the face, do you expect to catch someone running through an airport, or just walking down an office hallway? What are the lighting conditions?

Generally, the faster fps and shutter (1/30 or higher) are going to give you better results. The problem is that fast shutters in low lighting conditions can render your image completely useless.
I can't imagine fragmentation really being an issue or concern for a VMS. Fragmentation is a function of an operating system utilizing a storage device, and while generally, any NTFS system is going to try to write without fragmenting, as soon as the disc approaches a state where it's relatively full (70-80%) it's going to fragment the data to fit it on the drive. I don't know of any way to avoid fragmentation.

Either RAID5 or RAID6 will be recommended depending on the installation environment. High Security, Casinos, Health Care may desire the extra flexibility of having 2 drives fail without losing any data. Another option is RAID5 with a spare, but even then if a second drive fails before the spare is fully initialized, you're going to lose data. With that extra backup, it's going to cost more. Other customers or environments maybe fine with monitoring if a drive fails, and not paying extra for RAID 6 or 5+spare. Some IP NVR solutions also may only support RAID5 and not RAID6.