Establishing my own mail server

I will write more detail on what I did later.

For my own records, I relied on:

This article to set up postfix, dovecot, amavis and spamassasin.

https://www.exratione.com/2014/05/a-mailserver-on-ubuntu-1404-postfix-dovecot-mysql/

I removed clamav for performance reasons.

I implemented roundcube, but found it lacking, and replaced it with the much more attractive AfterLogic Webmail-Lite

http://www.afterlogic.org/webmail-lite

The Basics

One of the first things I needed to establish was a base of operations. I am lucky that where I live I have access to fast fibre internet.

So my options were:

1. Host at home

Pros, basically no cost, full control
Cons, not sure how this would impact on other home internet services like streaming, dynamic IPs, not using static IPs and having my IP readily identified, uptime unclear

2. Host out of home

Pros, uptime, disconnected from my physical location
Cons, cost, connected to my credit card

In my first iteration I spent a number of months working with Ubuntu on an HP Microserver using VMWare EXSi.

What I discovered was:

  • Although I haven’t done any professional nix work I was able to pick up linux very quickly based on a IT degree a decade ago and my wits and perseverance.
  • Running a VMware server is a unnecessary pain and I ditched it for Ubuntu 14.04LTS direct.
  • Dynamic IPs don’t seem to cause as much of a problem as you might think. I used DNS to bind to a dynamic DNS service with a reasonable level of reliability.
  • Internet uptime was not an issue on Fibre.
  • The major uptime issue was power outages. By the time I looked at a UPS or some other form of continuous power I was looking at needing to spend over $100.

Given I knew I’d be up for some cash I made the call to look at a cloud based VPS. *HYPOCRISY ALERT*. Did I just say Cloud Based? Yes. Basically I wanted something online, full-time, that I controlled, or at least could secure to the point where whoever was holding it could not access my data.

My home based server hasn’t gone to waste. I now have a very efficient media server, personal cloud server (to access my media when I am away from home), as well as a remote desktop that I can use to download and route information as required.

So three months later I ran up a Single Core, 512mb SSD based VPS with a fresh top level domain.

Next – considering security

Leaving the cloud

Dear Cloud,

I am leaving you. It’s not me, its you.

After watching Terms and Conditions May Apply one night on Netflix, I became interested in the amount of my own personal details were available ‘to the cloud’.

What is the cloud? in my view ‘the cloud’ itself is more a marketing term than an technical one. For the purposes of this exercise I am defining ‘the cloud’ as entities that hold my personal information in a way that is out of my personal control.

What am I on about? Just like the next person I rely on digital storage and exchange of information, and don’t intend to stop. The point I think I want to make is that services like email and social media existed long before it was called ‘the cloud’, yet its only in the last ten or so years that, at least for our personal data, we have come to depend on such a small number of ubiquitous free or low cost services like gmail, facebook, ebay and paypal. Moreover, its only recently that its become quite apparent as to the extent these organisation can use the data they collect to deliver those services in ways we don’t understand.

What I am concerned with is how much of that information leaves my control. So what I am examining is not the going off the digital grid, but more just limiting the amount of identifying and personal data that I may not be able to reclaim later.

I will aim to find real alternatives for:

  • routing a great chunk of intimate conversation through gmail  – so I will need a fast and easy to use mail engine with similar features
  • google contacts and google calendars so I will need a tool that can still sync with my devices
  • identifying myself on social media - I want to disconnect my social footprint from my identify, if thats even possible
  • using common identifiers in ubiquitous services I cant avoid - so that paypal doesn’t know me to be the same person who posts on whirlpool

But why? I have been asked by quite a few people, including my wife, who is mostly concerned that she won’t be able to tag me in facebook, and my tech qualified friends, even the more extreme of whom think I have become a whack job. There are a few reasons.

  1. I want to see if its possible
  2. I think it might come to be important – it is not out of the realm of possibility that one of the companies already mentioned may be required under law some time soon to keep all of this data indefinitely, and I am not sure I am ok with that.
  3. I think other people might be interested.

This is not me saying that I will never give up on all of this and go back to using the same systems we all do. After all, it is convenient. But I figure whilst I have the interest and the time (well, kind of) I may as well give it a go.

Roland Johnson (not my real name).