Back in the very old days of email, running an email server was a serious challenge for many reasons, like the cost of hardware, the cost of software, the cost of bandwidth, the cost of support services, and the cost of securing email and fighting spam. At that time big tech recognized the opportunity to provide easy to use email services for small companies and small communities, and thus Google started Google apps, and Microsoft started selling Exchange in the form of software as a service.
That situation has changed a lot in the recent years, Hardware costs are now very reasonable, especially using the AMD Ryzen or Epyc platforms. Bandwidth costs are also a fraction of what we used to pay 15 years ago. And While spam is still a problem, there are very effective automated solutions to secure user’s mailboxes. This is an opportunity for small companies and communities to run their own email servers, and own their email and their data.
I understand that the market is going in the “Email as a service” direction, and service providers promote their services using many nonsense TCO calculations trying to prove that they are cost-efficient. Communities that value their privacy will not be fooled by such marketing techniques, and can realize the benefits of owning their own data while nobody can spy on their users.
However, if you ask your geek friend about setting your own mail server, there is a probability that you will get an answer like: “Why, we have free and powerful email providers.” or “Do you want to go back in time? Nobody does that anymore”. I Disagree, and here is why.
What you lose
When you surrender your community’s email service to a big tech like Google or Microsoft, you will be giving them a tool to profile each and every member of the community. Things like “interests” and “known associates” can easily be extracted from each person’s email and sold separately or as part of advertising campaigns that target your users personally. If you are a school or a university, you will be allowing your big tech provider to data-mine your student’s essays, their thinking, your teachers’ thinking, and their interactions. Such information can be easily used in a devide-and-conquer campaign against your community, or in targeted blackmail against your users.
If you think you are too small or too insignificant to come on the radar of a big tech malicious campaign, remember that big tech corporations are run by humans, and they employ humans. Those people will have access to a wealth of information about your community, and you never know when or how one of them will be ready to sell access to your data or to target somebody you care about.
Therefore, Although your current big tech email service might be working fine for you, you have an interest to move to a privately hosted mail service to increase your security, and mitigate a wide variety of unknown risks.
Is it easy?
Although the setup process can be automated to a high degree, running an email service is not easy, but neither is fighting divide-and-conquer campaigns. It would be relatively easier if you are starting a new mail service, not migrating an existing service. However, no matter how hard it is in your context, it will be worth the effort.
It would also be relatively easier if you calculate your capacity based on your active community, and limit signing up to the service based on invites or admin approvals. After all, if you plan to serve up to 1000 users in the next 5 years, you can get on with a simple dedicated server. More users will require more space and more processing power.
If you have more than 1000 users, I would suggest that you decentralize your mail service. You can do that by creating multiple mail servers for parts of your community. Although some tech guys might advise against that, I think the benefits of decentralization are worth the extra investment in hardware and maintenance.
Focusing on up to 1000 users per mail server makes the task relatively easy. A single system administrator from your community can take care of the mail service, and run it smoothly.
What you need
If you have access to a system administrator, you would need a computer to act as your mail server. A computer in your closet can work in some special cases, but there are known problems for hosting email servers at home, therefore, for better stability and security, it would be better to host the computer at a data center, or to rent a dedicated server from a hosting provider.
Once the server is online, you would need the mail server software. The easiest way to start is to use “Mail-in-a-Box” from https://mailinabox.email. This is a free and open source tool that prepares a Linux server to become a full featured email server. The tool has a video on its home page that shows in details all the steps you need to take to have a running mail service.
What you get
If you follow the installation guide of Mail-in-a-Box, your mail service will be up and running within 1 hour. And in addition to the mail service, you will be ungoogling your community’s calendar, contact list, and file sharing services just by using the features of your server.
Users who use big tech email services might find the user interface of your mail server a bit different, and that’s normal. Microsoft and Google web mail services are designed for tracking their users and optimizing their advertising revenue, and unfortunately, Users became familiar with those interfaces against their own benefits.
Running your community's mail server is a decentralization initiative, and provides all the benefits of decentralization. In addition to that, it deprives google from some data about your users, and with more decentralization initiatives like running your own search engine, running your own community social network, running your own community video streaming service, among others, you will be reducing the impact of big tech’s divide-and-conquer risk.
Are you ready?
The ideas above are open for discussion with decentralization advocates. If you are interested in a discussion about email hosting please leave a comment below.
The Unweb Developer