We discontinued our VPN services mainly because of technical issues (there were also some bugs in the billing module but these weren't crippling most of the time). Basically, multiple times a month we would have some clients that couldn't connect to some of the VPN servers for whatever reason. I would spend a few hours rebuilding the RADIUS databases, then rebuild the RADIUS servers (3x), then rebuild some of the VPN servers (sometimes all 15), and then all users could connect to all VPN servers again. This week, we had an issue where all but one account could only connect to our FL VPN server (OpenVPN and PPTP) and none of the other ones. The one account that was the odd man out could connect to 2 other VPN servers but not the rest. I built brand new VPN servers and added them to the pool and still no dice for any of the accounts (including the 1 odd account). I rebuilt everything from scratch and I couldn't even get the authentication servers to see attempts from the other accounts.
Not including the VPN servers themselves (luckily we only owned the hardware for 4 of the servers), we've spend hundreds of dollars on our CakeVPN infrastructure and in the end it was a lack of knowledge on my part and a few painful bugs in the management software that were the deciding factors on how to proceed. Yes, our VPN clients made up less than 0.1% of our total clients, but for those clients who still want a VPN in FL, CO, CA, or IL a 32MB VPS is much cheaper and a worry-free solution.
Since we'll be keeping all of the software and the licenses we hope to reopen CakeVPN in the future when we get a VPN expert on staff who can maintain the servers and backend properly.
I apologize for those impacted by this decision and I am more disappointed than most because of the time I invested into this hoping to have something great to share with others. Even after the e-mail was sent out I spent 3 hours this morning trying to figure out what the problem was and how to fix it in hopes of learning how to fix it in the future but it was 3 hours wasted.
I would also like to add that we created these VPN services to meet a demand that we had potential clients hounding us for, we had A LOT of pre-sales tickets from people who did not want to manage or maintain their own VPN servers so we decided to build CakeVPN. Surprisingly, after we launched it the interest died off immediately. For the first couple of months it was smooth sailing, the technical issues only popped up in the past few months and this last issue was the straw that broke the camel's back. It's a shame too because aside from the initial investment, the operating costs were very minimal but the time invested was just to much for me at this point in time.
