Artificial Intelligence and Grid Computing
Scott Carnahan
Net 204-102

Artificial Intelligence and Grid Computing
This paper is a look into how Artificial Intelligence (AI) can be used in a network environment with grid computing. I will briefly look into how AI works, grid computing and what you get when you put the two together.
What is Artificial Intelligence and how does it work
Artificial Intellegence is a branch of computer science that atemps to simulate intelligent behavior in computers, that is to say, to get a machine to imitate intelligent human behavior(Merriam-Webster) and with any luck ignore unintelligent human beahviour. AIs have to learn concepts of the things they observer or work with, they then store what they have learned, like any other software, as chunks of data. An AI is usually presented with a goal to achieve they accomplish this by exploiting the relationship between the data they acquired and the goal they are trying to achieve (wiseGEEK). To put it into a simpler concept; AI is broken down into an If, and Then type programing analogy.” If this happens Then do that”, pretty much like Basic Programing. But the If-Thens are attached to something not so simple, complex algorithms that are used to record what all the results of the Ifs and Thens are, so that the data can be accessed by the AI, then when the condition happens again it knows what to do. For a little better explanation of AI algorithms look at Google's cloud-based machine learning tools, they use learning algorithms to analyze data. A Google algorithm is added to the logic of a computer or “machine” the results are stored in a dataset, which in turn linked to a larger, smarter dataset. The algorithms learn by recording historical click data, social data, and link data. Then the datasets from algorithms are added to the master equations. A master equation is an equation which describes the evolution of the probability of a state at a given time as the balance between transitions leading to that state and transitions removing the system from that state (Dictionary.com). In other words the master equation is a complex set of algorithms that are used to predict the probability of what is going to happen next. That’s how Google makes their money, they learn the kinds of things you look at, buy, the people you talk to, the websites you visit, they match that with what historically you do when you are online and what other people with similar interests do, and then thanks to their “Master Equation” they know what ads to put around your searches and what you are going to be looking at next (Google Prediction API).
Grid Computing Red Lambda With the basic idea of AI, apply that to a network with the concept of grid computing. Grid computing is when the unused CPU are shared across a network using a parallel processing architecture, all the participating computers allocate their resources to an application that is very computation intense, in essence making them a super computer (PC Magazine ).
A company called Red Lambda has combined grid computing, with break troughs in artificial intelligence, and put them together into a dashboard type control center called MetaGrid. MetaGrid software is billed as an automated security discovery application for organizations with what the firm calls big data IT, network and cloud infrastructures. “What this software does is relatively unique in the security marketplace – it first looks at the data that is available locally and then allies this with the remote information. It then collates and analyzes all the information, then presents to the user as information feed” (Krautkremer, 2011).

MetaGrid Takes Advantage of breakthroughs in information science. Applying MetaGrid capabilities to massive data sets, then arranges events into clusters that all share common patterns. Each cluster is a compressed representation that makes it easier to sort through the data. At the same time MetaGrid catalogs anomaly events based on what it has learned in the past. Then it makes a visual representation of the most the most unusual events along with their operational content (Red Lamba).
It is not an easy task to make sense of large amounts of data, most of the time data is simply archived and indexed without much else being done to it, data is often confusing and hard to analyze on a large scale. Using the breakthroughs and the idea of grid computing, MetaGrid can find patterns in data and see things that you might not have been unaware of before, by using the following using three tools (Red Lamba). 1. Clustering—Simplifies data and organizes it, making it smaller and easier to handle without losing detail. 2. Classification— MetaGrid learns from operators to tell you what is based on what it has learned, and attempts to predict when these events will reoccur. 3. Anomaly Detection— It looks for patterns, events and timelines that are unusual.
The old way to correlate data is to run manual rules over it. MetaGrid looks for relationships in all the data is correlates, which is automatically done by its AI engine. It can work with many sources at the same time, without using manual rules, giving you a better understanding of how data flows and reacts with your network. MetaGrid AI also data from outside data feed through 3rd parties that are in context with local data processes. These additional data sources provide more information to the data being processed locally on the MetaGrid (Red Lamba).
By combining data feeds, and correlating them with its AI, all data in the grid is analyzed, classified and compressed. It then overlays the data from the different contexts to create a workflow for the operators working with the network, this gives them the ability to mitigate situations, resolve problems and avoid threats in a more timely manner (Red Lamba).
MetaGrid’s AI can give immediate response to threats, firewall breaches, and malware infections that might affect data, systems or infrastructure. MetaGrid has a scalable policy engine that can push mitigation through the network, and it can do this on scale that until now, has been unprecedented (Red Lamba).
MetaGrid Search lets you look through data after it is indexed and processed by the MetaGrid platform. Combined with correlation the data is then presented with context.

Red Lambda uses a “transitional co-existence” approach, meaning it allows you to plug in existing systems into it; this allows you to extend the life of your investments, while adding a full suite of advanced analytics, business intelligence and operational solutions and is capable of integrating them with your existing solutions (Red Lamba).
The MetaGrid Unify API gives you two integration points: one for control through MetaGrid and one for 3rd party apps. The MetaGrid Web API lets you manage MetaGrid’s platforms in real-time, providing visual information of the network loads and simultaneous jobs that are running.
The MetaGrid Services API provides the ability to integrate your existing applications into MetaGrid as services (Red Lamba).

Conclusion Red Lambdas MetaGrid is a huge step forward in AI in a network environment. From what I can tell, it is just coming out of its beta stages. It also seems to be geared for huge data centers and networks at the moment. I’m sure scaled down versions are on their way and I am sure other companies are working on similar projects. I found several research papers on using AI in automation, security and network deployment, but they were way over my head.
One thing I didn’t touch on is that current OS’s and applications use AI already. When a useful AI matures and is stable it is put into the back ground of operating systems, applications like office suits or anti-virus programs. AIs are all around us, we use them every day, and we don’t even realize it. They make our smart phones more intuitive, they make are cars feel a little smoother to drive, and they are getting better all the time.

Works Cited
Dictionary.com. (n.d.). Retrieved 11 5, 2012, from Dictionary.com: http://dictionary.reference.com/browse/master+equation
Google Prediction API. (n.d.). Retrieved 11 5, 2012, from Google Prediction API: https://developers.google.com/prediction/
Krautkremer, T. (2011, 9 29). Todd Krautkremer.
Merriam-Webster. (n.d.). Retrieved 11 5, 2012, from Merriam-Webster Dictionary: http://www.merriam-webster.com/dictionary/artificial%20intelligence
PC Magazine . (n.d.). Retrieved 11 5, 2012, from PCmag.com: http://www.pcmag.com/encyclopedia_term/0,1237,t=grid+computing&i=43962,00.asp
Red Lamba. (n.d.). Retrieved 11 7, 2012, from Red Lamba.
Securebuzz.ca. (n.d.). Retrieved 11 6, 2012, from Securebuzz.ca: http://securebuzz.ca/2012/07/ai-and-grid-computing-disrupting-traditional-network-security-213/ wiseGEEK. (n.d.). wiseGEEK. Retrieved 11 5, 2012, from wiseGEEK: http://www.wisegeek.com/how-does-artificial-intelligence-work.htm