Hosted services vs. self-hosting
I recently had a Kerio Connect customer ask about hosted email. Their goal was to have better resilience against natural disasters - they don't want local problems to kill their email.
You'll find plenty of people recommending this. Some of the arguments make good sense - for example, if you go with a big company (Google or Microsoft, among others) and they have problems in the data center where your email happens to live, they can often quickly shift you to another location and you may experience nothing more than a hiccup.
That's not necessarily a good thing. If you are a one location company and your local Internet access is down for whatever reason, which is better:
- People can send you email but you cannot read or respond to it
- People get an error when they try to send email
In the first case, your customers don't know why you are not responding. In the second case, they do know you have a problem. Shrug.. neither situation is ideal.
Other arguments are equally ambivalent. While some companies might want to shift maintenance and troubleshooting tasks elsewhere, others most definitely want to retain control. In some cases (Google is perhaps a good example), it may seem that you give up a lot and don't gain much because most of the maintenance tasks are still yours. In other cases, you give up everything and can't even add a new user yourself.
Even when you do have access to logs for troubleshooting, delays can be maddening. Google tells you that it may take up to an hour for information to appear in the logs. In my testing, it took about 45 minutes on average, but that delay could be very annoying:
Cost is another issue. Only the very small companies can actually save money by buying email hosting. Some years ago I wrote "Hosted Google Apps Gmail vs. in-house Mail Server" and felt that the break even point was around 10 users. It's less than that today because computer equipment cost is less, but it's also true that companies with less than 10 users are often willing to go with free email hosting.
There are other arguments (from Self Hosted Mail Servers Just Don’t Measure Up):
I think that's quite exaggerated. Most of that is things every company needs to do anyway, so adding a mailserver to the mix is hardly the gargantuan task implied there.
Still, that disaster recovery stuff is important. Having your email in the cloud definitely has advantages. However, you can get that advantage without losing anything.
Rent raw space, not your email
That's what I pointed out to my customer: you can get the advantages of email hosting by renting raw server space, installing an appropriate operating system, and then transfer your existing Kerio Connect to it.
That's what I do. My mailserver runs on a Linode server. I also run my webserver and Kerio Workspace there and I'm no where near exhausting that server's capacity. The cost? $30 a month.
If I remember correctly, I started out in a rack somewhere in Georgia and I'm now somewhere in California. I forget why they moved me - maybe there was some problem at the first facility or maybe it was when I upgraded to a bigger server than I first rented. The point is that this company is big enough to have multiple locations. If my California hosting falls into the ocean, they can relocate my server somewhere else - even Asia if I wanted it there!
I get the disaster recovery advantage, I'm not maintaining hardware, they have UPS equipment, redundant Internet access, backups and all the rest, but I still have full control of my email and am not paying outrageous email hosting fees (which run from $4 per user per month on up, depending on services and disk space desired).
For me, that's a good situation. I don't need a dedicated server (Linode is shared Xen) and honestly most email users can do fine with shared hosts - email is a relatively light load that requires little in the way of resources. You'd need to have a lot of very active email users to need dedicated hardware.
As I am a very small company, needing email only for myself and my wife, I actually could go with free or low cost email hosting. I'd give up control, though. I wouldn't be able to debug delivery problems, install new versions on my schedule, add my own scripts to process email automatically and so on. Those things are important to me, so the minor cost savings are meaningless.
Got something to add? Send me email.
(OLDER) <- More Stuff -> (NEWER) (NEWEST)
Printer Friendly Version
Increase ad revenue 50-250% with Ezoic