heironymouscoward's Journal: Last, ever, journal entry

This is my last ever journal entry.

HeironymousCoward is signing off.

The news today that criminal gangs have been spotted selling botnets, the day after the FCC considers spyware to be a valid model for software producers...

In case there is any doubt at all, let me summarize the near furture of the Internet in two words: Robber. Barons.

The installed base of computers and their hooks into modern society represents an incredible resource. You can look at crime from many angles, but the one that I prefer is the biological/ecological metaphor of the parasite. For every "normal" species, there are 3-4 parasite species. The evolution of life has been largely driven by the need to fight off opportunistic parasites.

The Internet, representing the infrastructure on which the modern information economy is hosted, has almost no defenses against parasites. Large swathes of it are monocultures, and the number of successful invasions of this monoculture is on an upward curve that shows no sign of slowing its rate of increase.

Historians know that the present is never framed in terms of arguments of the past. The future is never framed in terms of arguments of the present.

We look today at the Internet as a vast market opportunity, and debate how companies like Microsoft can be allowed to exploit this market without harming it. We argue about the relative merits of alternatives. We predict their growth and discuss strategies.

It's all irrelevant. Tomorrow's Internet will be concerned with only one thing: fighting the war against the invasion of the body snatchers, the infinite armies of parasites that infest every single susceptible computer.

Let me make this concrete. You want to use the Internet? Start by paying a small fee to your local security service provider. And hope he does his job. If his protection does not keep you safe from the others, your network access and data will vanish randomly. The good news is that someone, somewhere will be back with your data, for a price.

It will take only a generation for common criminal behaviour to turn into formalized protection, and from there into a form of taxation on all commerce.

We could sit back and watch this happen, except that during this process, governments will not sit still. Huge task forces will be assembled to protect the national interests and fight the criminal gangs. Wish it was that easy. For every success, ten new gangs will spring up, more vicious, more creative. The anti cybercrime task forces will spend huge amounts of money on the fight, will lobby for extra powers, will use increasing force and aggression.

Caught between the increasingly organized and motivated criminal gangs, and the symetrically developing cybercrime task forces, will be the citizenry of the world, as usual. We will watch as our liberties are infringed from both sides. We will complain but no-one will listen. And we will pay, one way or another.

By the time the war is over, after 20-30 years or so, there will be little difference between the forces of law-and-order and those of the underworld. The rule will be: if you want to do business, you pay. If you can't pay, go and work for someone else.

That is what I predict will happen.

Now, there is a simple and obvious way to prevent this. I'll let you figure out what that way is, and why it's not going to be.

Personally, I'm retiring to start a microbrewery.

Hasta la vista, friends.

I don't think so.

[rdewald]

I understand your outlook, but I think you underestimate the forces working against these problems. My personal theory about what's needed in the short run is a new model fo OS architecture.

For example, I am working on a project that I can't talk too much about in public, but the concept is this: workstation terminals boot over the network from a central image which is maintained vociferously with multiple itegrity checks. No OS local to the machine, just flash storage which is relayed from another ser

Re:I don't think so.

[heironymouscoward]

My company built systems like this in 2001-02, using network booting, for a kiosk system. The kiosk boots, uses dhcp and bootp to find a boot server, tftp to load its OS, tftp again to load an application set, and then dhcp to find an application server to connect to. Yes, it makes very robust clients - the boot server is itself booted purely from CD-ROM. It was great fun, I designed the hardware myself, and the kiosks have been running in a production environment (cement factory) for two years without a

Re:I don't think so.

[Sloppy]

Imagine the following scenario: using Windows botnets and Darwinistic programming techniques to evolve parasites capable of attacking 'secure' systems. I.e. create a bunch of attacking algorithms, provide them with the rules to compete and breed, and use the infected population to power the whole thing.

...
I believe it is a matter of time (20 years at most) before _any_ computer using hand-written software is de-facto compromised within seconds of being connected to the Internet.

No way.

It's possible t

OpenBSD as an ecology

[heironymouscoward]

You are right to say that without a degree of success in every generation, no evolution is possible.

The ecology in which parasitical software lives has many levels. There is the network itself: the TCP/IP packets, routers, hubs. There is the OS. There are the applications. And there are the users themselves.

"Success" for a parasite does not require root access on an OpenBSD box. Imagine that some exploit evolves or is designed to break into a popular web browser or email client. All OpenBSD boxes ru

HC

[Raindance]

I'll be sorry to see you go.

Will you be keeping a blog or some such, while microbrewing?