Dedicated Server And ARP Poisoning

I recently had a problem with a hacked dedicated server which was attacked by ARP Poisoning and a Remote Desktop man-in-the-middle attack from another dedicated server on the same subnet. Maybe unreasonably I expected controls in place to prevent this, better detection and better handling of this problem, lack of which have left me uneasy about the hosting.

I know using Remote Desktop with a cert would prevent the server being compromised, but my concern would then be HTTP traffic being hijacked and malware insertion, redirection to non HTTPS login pages, redirects to anywhere, etc. If ARP Poisoning occurs then even if my server is fully secure all the web addresses pointing to my server's IP are basically compromised by HTTP traffic redirection.

Before this happened I had assumed (bad idea) that there would be some kind of mac level assigning of IP addresses.

What level of protection from this type of problem should I expect from the Dedicated Server supplier on their network? Problem started after I rebooted our server, IP was grabbed and the network adaptor was disabled due to IP conflict, so machine didn't not respond to pings. I raised a ticket and was told

"when your server came up it couldn't use it's assigned IP address as for some unknown reason another device on the network is using it's IP, we're tracking down the device and we'll have your server operational in few minutes."

They re-enabled the network adaptor presumably without fully checking the situation. I assumed the situation was either an innocent misconfiguration or that the issue had been fully investigated and dealt with, I reconnected via remote desktop and a few minutes later the server was compromised (wiped event logs, Cain and Abel installed etc).

Our machine was wiped, reinstalled and no further problem arose, but they initially seemed to deny that the two issues were related. Suggesting it would have been hacked externally via IIS vulnerabilities. Then 18 days later(!) they released a message advising all users with machines on the subnet that they had shutdown a malicious machine (not ours) on the subnet and to change passwords, run malware scans etc! Whether this was the same original machine or another server compromised I don't know. However our server was running with Cain and Abel and a whole lot more for quite a while as I checked it before it was taken off line for reinstallation.

Is this a common occurrence? Do most dedicated hosting providers have proper measures to prevent this or are there any measures I can take to prevent this happening again?