Applications as Virtual States

This is an excerpt from my article Building Super Scalable Systems: Blade Runner Meets Autonomic Computing in the Ambient Cloud.

As I was writing an article on the architecture of the Storm Botnet, I couldn't help but notice the deep similarity of how Storm works and changes we're seeing in the evolution of political systems. In particular, the rise of the virtual-state. As crazy as this may sound, I think this is also the direction applications will need follow to survive in a complex world of billions of compute devices.

You may have already heard of virtual corporations.  Virtual corporations are companies with limited office space, a distributed workforce, and production facilities located wherever it is profitable to locate them. The idea is to stay lean and compete using the rapid development and introduction of new products into high value-added markets. If you spot a market opportunity with a small time window, building your own factories and hiring and engineering team simply isn't an option. Building factories is a bit old fashioned and is left to the select few. These days you get an idea for a product and contract out everything else you possibly can. It doesn't really matter where you are located or where any of your partners are located. If part of your product requires a specialized microprocessor, for example, you'll contract out the R&D and the design. The manufacture will be contracted out to a virtual fab, then the chip will be sent to a contract manufacturing service for integration. Look ma, no hands.

Futurists say land doesn't matter anymore. Nations don't matter anymore. Entire relationships are abstractly  represented by flows of money, contracts, information, and products between all these different agents. Interestingly enough, what technology is the absolute master of managing flows? Applications! But we are getting ahead of ourselves here.

Let's first get a definition of virtual state. Foreign Affairs magazine published an article titled The Rise of the Virtual State, which defines virtual state as: a country whose economy is reliant on mobile factors of production.

Immediately it looks as if only old style countries can be virtual states. That's really just a historical bias. A quaint throwback to when land was king. R Jagannathan sums it up nicely with this quote: In the post-modern, post-internet world, many states exist more in the mind than in reality.

So let's just say a virtual state can be any entity requiring political representation. Surprisingly, software applications fit this definition. Applications deal with high finance, they negotiate, they require global economic access, they execute policy, and they operate globally in many legal and political jurisdictions. If that kind of entity doesn't require political representation then what does? And after all, if you are going virtual, why not go all the way?

The mobile factors of production part of the definition is a bit confusing, but turns out to be a key driver for why applications make sense as virtual states. Factors of production are the resources employed to produce goods and services, but do not become part of the product.

Historically these factors have been land, labor, and capital goods. And historically agricultural output has been the biggest component of GDP. Agriculture relies on immobile factors of production like land, climate and geography. If you want to grow a crop it will be difficult to compete with someone if they have better land and climate than you do. With their immobile factors of production they have a competitive advantage. For example, I live in the under appreciated Santa Cruz Mountains wine appellation. The wines of this region reflect the dramatic microclimates created by a unique mixture of mountain topology and ocean proximity. Nobody else in the world can recreate this sublime mix of circumstances. However, advantages of this sort have become less and less important in the evolving modern economy.

In the modern era, we've added a few more factors of production: financial capital, technology, human capital, intellectual capital, and business know-how. The big difference in this list when compared with the original list, is that these are mobile factors of production. They are mobile because they are abstract and intangible, they aren't tied to land, they can be anywhere in the world. Money flows. People flow. Knowledge flows. Contracts flow. Skills flow. And trade flows to the entities with the greatest absolute advantage in these factors. If, for example, you land a contract to represent the next great NBA superstar directly out of grade school, you think to yourself why pay Nike all that money to make a sports shoe? You can hire all the experts (design, finance, manufacturing, legal, marketing, etc) and keep all the profit. It doesn't matter where these experts are as long as they can all communicate. They key to the deal is the marketable element, your NBA star, once you have that you have what matters, the rest can be assembled.

They key takeaways from this analysis are that in the modern economy land isn't key. Making money depends on intangible and abstract skills that can be located anywhere and do not require a physical presence to carry out. The ability to create and manage these flows between networks has become a prime factor of production.

Land not being primary means countries are no longer in the business of acquiring more land, which is quite a change historically as pretty much all of human history has been about killing people to acquire more land. The prime example of this is Hong Kong. A measly 426 square miles in size, Honk Kong has become a major economic player on the world stage.

The emancipation of statehood from land makes for a very confusing world. One consequence I see is that a virtual state is really not different than a virtual corporation, there's simply an added political dimension attached. The political dimension stems largely from historical land entanglements in the form of nation states. To any observer of the political process it's quite clear corporations have their own political interests, so the distinction quickly becomes moot, which means the number of virtual states will multiply radically. John Robb even argues that Al Qaeda network is a sort of virtual state, with a consistent source of finance, a recognized hierarchy of officials, foreign alliances, an army, published laws, even a rudimentary welfare system. It has declared war on the U.S. for much the same reason that Japan did in 1941: because we appear to frustrate its ambitions to regional hegemony.

The goals of the state and corporations have also merged. The goal of a virtual state is to make money, not acquire land, rule people, or seek a greater good. It's entirely selfish in nature. A lizard mind with the Internet as its body.  There's no centralized bureaucracy and no official armed forces. The role of the virtual state is to develop an overall strategy;  invest in people, not expensive production capacity; contract out other functions to states that specialize in or need them. Pretty much exactly the same mandate as a virtual corporation.

Why does this matter? This is a question I try to ask myself before I go to far off the deep end. Well, I don't think applications will be virtual states in the sense of independent AIs and all that. This dystopia is often portrayed in SciFi Novels, but I'm really trying to picture how this system will work, not just make up a story.

To really make this work humans and applications will form close partnerships and it's that unit that becomes a virtual state. Multi-national corporations have gone way past caring about their association with a nation state, this trend will continue down to lower and lower levels of organizations. The automation process that has always been the strength of software makes this possible. With the hollowing out of the middle class each of us will be left to make a living in this new environment and applications will help make this possible.

Applications can not be autonomous because software doesn't have goals or objectives. Only humans have goals and these come from a long history of physical and cultural evolution. These goals are imprinted into applications and it's the applications job to achieve the state objectives.

We've already seen how applications play this part of the partnership with black box trading in financial markets. These black boxes are programmed with goals and models and tactics by a very select group of people. Few others have any idea how these things work or why the computers decide what they decide.

Another stunning example of how this people-computer partnership would work is detailed in the Wired article The Answer Factory: Demand Media and the Fast, Disposable, and Profitable as Hell Media Model, written by Daniel Roth. The story is about a company called Demand Media. Demand Media data mines the topics people are searching for on the web using a complex monitoring process and algorithmic sifting. For each candidate topic it's calculated what advertisers are willing to pay for hits on the topic. Profitable topics are then put into play. Someone is hired through an automated bidding process to create the content at a very low price, low enough to still be profitable given low adsense payouts, so these people are working for cheap. Once produced the content is published and the adsense rewards harvested.

This is an amazingly sophisticated operation that is almost completely driven, coordinated, and managed by software. It's not hard to project this strategy out for material products as well. Software can scour the web, data mining potential products. Information from human trend spotters can be integrated in. The determination of what products to make can be made through some combination of automated and manual processes. After that much of the rest of the process can be automated. Financing can be arrange through some sort of automated credit check. Building and designing the products can sub contracted out. Manufactures can be fed the results of the design. The product can be automatically made available retail outlets, put on Amazon, put on eBay, put in vending machines, made available to affiliate programs, and so on. Commercials can be contracted out for shopping channels. Firms can be hired to create advertising and marketing campaigns. Time can be automatically purchased on radio and TV stations. Bloggers can be bought to write articles. Social media experts can work the microblogs. Reviews can be added to Amazon and Yelp. People can buy on-line or at a mall or wherever it's convenient. The accounting, reporting, and billing is automated too. All the while the application is learning using measurable feedback so it can do a better job next time.

And consider: almost this entire process can be automated and controlled by an application. It would operate relentlessly, 24 hours a day of everyday, churning out products and services. This is actually a damn scary thought.

The key is to recognize it's all about automated and structured flows controlling mobile factors of production. Software rules processing these kind of flows. It can do them faster, smarter, and it never gets tired. All that is required is an entire chain willing to work on contract, carry out an objective, and work in a coordinated fashion.

In this way applications become virtual states. Again, not in some silly SkyNet sort of way, but in a very practical this is how applications will need to be structured in order to work across billions of devices sort of way. Software has traditionally been thought of something contained, something within well known and controlled boundaries. The boundary was a single machine at first, then small clusters of machines, then huge clusers of machines, then even entire clouds. Crossing the Ambient Cloud boundary means that software is not contained anymore, software does the containing. It moves where it needs to for reasons dictated by its high level goals. It expands and contracts as needed. It negotiates with 3rd parties. It protects itself from attack. It protects itself from failure. To a great extent software becomes independent, thus the idea of software becoming like a virtual state.

The criminal networks behind botnets like the Storm Botnet have already made the transition to a virtual-state, operating a sort of black hat bazaar that has both  economic means and military capability.

For a good overview of the black market economy please read The Cybercrime Blackmarket. It documents an amazingly rich world that doesn't physically exist in any one place, yet is united behind common objectives and methods.

In the cybercriminal black market we have leadership by the criminal heads, a highly skilled and motiviated work force, a means of making money, a currency, and even a loose system of laws and morality.

And oh, they also have a military. Not a military arm not in the conventional sense, but the same DDoS capabilities that are sold on the market can be used as a weapon. Blue Security was literally forced of out business by a relentless DDoS attack carried out by an organized crime ring angry at the success of the firms anti-spam software. Other digital and physical attack strategies are certainly within their grasp.

It's telling that the same group that first learned how to exploit the Ambient Cloud is also one of the first to form a virtual state. The two go hand-in-hand.

If you would like to read the rest of the article please take a look at Building Super Scalable Systems: Blade Runner Meets Autonomic Computing in the Ambient Cloud.