Rick van Rein

wo 19 november 2014


Back to Hosting

Over the years, the Internet has been neglecting its hosting branch a bit. Anyone capable of running their own server can achieve miracles, but those who cannot need to fallback on services provided. It is not surprising that these users landed with the few "free" service that reign today's Internet.

The Internet is constantly evolving, and new protocols and standards are being added each day by the Internet Engineering Task Force (IETF). The development of software keeps in pace with these developments, and a wealth of options for communication is available to us all. But in the end, that's not enough.

The original Internet

The most wonderful Internet imagineable is one where every user is in full control of their use of the Internet. Having software and a connection suffices to get online in any way one wants. Moreover, the software can connect to other software running one of those protocol standards that are constantly being developed. There is just one problem with this -- not all users have the knowledge or time to run their own servers.

In retrospect, it is no surprise that the initial surge in Internet popularity occurred together with the rise of hosting companies. Such companies help you to a domain name that represents your own corner on the Internet, and they run the servers for you, with possibilities to set it up as you desire. The so-called LAMP stack, consisting of Linux, Apache, MySQL and PHP was provided and users could go their own way.

The current problems on the Internet

As discussed in other postings on this blog LAMP has its own problems, but perhaps the most problematic problem of all is that nothing has been added ever since. PHP enables people to run all sorts of web applications, so when things like online payments and blogging are introduced, it is implemented as a PHP application. Meanwhile, the hosting industry came to a halt, and ended up competing with each other based on pretty exchangeable service packages.

The problem of the current situation is that all that a hosting deal supports is web and email. It's been like that for 25 years, or put differently, since the times that we all carried around floppy disks! And users simply want to get more out of the Internet -- and they should. No chat, no telephony (who wants to be a number? I don't!), no contact information available to clever client-side utilities that could dig it up automatically. All we have is the web and email. It's no wonder that more refined services have taken the lead on the Internet, in spite of their full control over user's online presence; users simply have little choice.

It would not be fair to put the blame on hosting companies, however. Things have simply grown this way. And a single hosting company is not able to affect changes; they need the others to follow suit, or else their work has been in vain. The introduction of new protocols has a very big critical-mass problem; this is much easier for the well-known silo's that basically create an internal domain, shielded off from the rest of the Internet. And it is only natural that developments start that way, and mature into the distributed model that keeps the Internet a robust whole.

The InternetWide project takes a very clear stand on how to get to this mature version, and thereby help users to regain control over their online identity. This solution consists of two things, described below.

A Hosting Distribution named ARPA2

One of our aims is to construct a hosting distribution that not just one, but many hosting providers can install. The distribution is a direct alternative to LAMP, and it introduces modern protocols alongside the ones that we already love. With the main expenses of hosting being bandwidth and maintenance staff, the cost of running such services is often low.

Being able to offer more abundant selections is helpful for hosting providers; even if they cannot charge much more for an added service, this scales up over the large numbers of domain names that they tend to host.

Note how this distribution overcomes part of the critical-mass problem, for one because all hosting providers receive the same offerings at the same time, and in part because a ready-to-go distribution simplifies the roll-out of those new alternatives. ARPA2 will take into account that users must be managed somehow, that they own domain names, and so on.

Specialising the Hosting market

New protocols can be knowledge-intensive. This means that not all domain hosters will be able to offer the full spectrum created by the IETF. This means that there is room for specialisation. Ideally however, specialisation should not lead to not running certain services because a hosting provider is not offering it. That would retain some of the critical-mass problem that blocks the introduction of new protocols!

To facilitate the optimal situation, there is a need for procuring a domain hosting package at one location, and plugging in externally hosted components from anywhere else. We therefore split the hosting landscape into two roles:

  • Domain hosters are those parties that register domain names, and offer a central cockpit to control it. Think of creating users, assigning rights to them, and... adding external plugin components.

  • Plugin providers run an individual service in such a way that it can be plugged under a domain hosted elsewhere, and even be controlled by it.

This model should automatically lead to a sincere payment model for the Internet. And since bandwidth is cheap and support can be a for-pay option, this should not lead to great expenses for end users. Actually paying for services used is fair, and directs money to the people who actually do the job. Moreover, as part of a contract the user is able to select those offerings from the market that employ reasonable privacy terms. Thanks to local laws in various countries, businesses are often forced to consider those privacy terms, and that ensures that the market will actually offer options with good privacy.

Go Top