11 essential algorithms that make networks work

August 11, 2011

11 essential algorithms that make the Internet work,” in Business Insider. This is a fascinating little slideshow—not just because it investigates the functioning of the Internet, but because with a little vision you can see these algorithms (or “functional processes,” let’s say) can be found in any network, any emergent system, and even in the general functioning of social relationships of life itself. Here they are:

  • Google Search: every emergent system—especially those that are larger—need some mechanism for finding something within the network. How often have you heard a YWAMer say the organization is so big it’s hard to know what anyone is doing? Better search is the answer. Part of search involves standardizing the presentation of resources so that they can be discovered (i.e. tags, or hashtags on Twitter, or the like), and part is in using a very good algorithm (or service). Making information easily findable via Google can help any organization or effort.
  • Routing algorithms: a core technology at the heart of the Internet which makes the Internet possible, yet one we never really think about. If we didn’t have the servers that tell us where to go to get what we want, we wouldn’t have the Internet. Directories are another form of routing. Lists of websites. Blogrolls. Linkedin Introductions. Highly connected “hub” individuals in swarms. Regional facilitators, national facilitators, team leaders in organizations. Major mission conferences. These are all ways in which we “route” to desired resources and opportunities.
  • Encryption. It makes credit card transactions possible, keeps email from prying eyes, and enables the safe sharing of information that could get people in trouble. Before 1996, encryption algorithms were considered munitions and sharing them was illegal. Today, without them, much of what happens in networks couldn’t happen. How much of missions has been enhanced by encryption systems? Encryption is the “force field” around some networks that enables them.
  • Facebook news feed. Sharing what’s going on is a key part of a network. But you also need a way to “bubble up to the top” what’s important. What’s important must be customizable down to the individual person: what is important to me may not be important to you. There are numerous arguments over algorithms for customizing the news feed because of this. Part of the challenge is that when I share something, I want you to see it—so some people try to “game the system” and beat the algorithm. News algorithm wars are constantly going on, especially on the top news feeds.
  • Sendmail. A basic system for sending a message from one person to another: not a semi-public news post sent to multiple people, but rather a private message sent to one or just a few people on a specific subject. Without the ability to send these kinds of subject-specific messages, collaboration can’t happen.
  • Netflix Recommendations. Ants do this all the time: they recommend a food source using scent. Another ant follows the scent trail, and lays more scent coming back. With every ant, the scent gets stronger. The stronger the scent, the more ants go to it. The power of recommendation based on what others like you say is good is critical for uncovering new resources. How do you know if it’s worth your time? Because two people who are like you have seen or read it, and say it is. Related: Amazon recommendations, Ebay recommendations.
  • Hashing algorithms. You may have never heard of a hash, but it serves a simple process: it authenticates the message and makes sure it hasn’t been tampered with. Encryption keeps it safe, the hash is the “signature” that confirms it really did come from the person who it purports to be from, and wasn’t altered along the way. Every network needs a way of knowing when a message is authentically from someone. The more digital communications can be faked, the more identity hacking will become a big issue in the future.
  • Advertising algorithms. We don’t necessarily like advertising—except when it leads us to something that is really useful. Like an advertisement for a new book that we end up loving, or for a conference we want to attend. Relevant ads are useful to people and can generate revenue. Networks need mechanisms for “classifieds.”
  • Google News. The Facebook news feed is where members of the network share their personal news—what’s going on in their own world, in their own interests. Google News, on the other hand, scans a far broader collection of news. A network needs the ability to know not just what’s happening inside their network, but outside as well. It’s not enough to know who’s doing what in Egypt within your network: Google News (or some similar algorithm) brings you what other, possibly competing, networks are also doing.
  • Amazon Recommendations. This was briefly mentioned above under Netflix Recommendations. The interesting thing about Amazon’s recommendations is that it’s largely by people who have purchased something, whereas Netflix is typically from people who have watched something. Amazon’s recommendations therefore involve financial purchases and whether it is value for the money, where Netflix is all about value for the time.
  • Pandora playlists. This is another form of recommendation. In this recommendation, the algorithm looks at the information that we regularly use—in this case, music that we regularly listen—and compares that to the information used by the people who are in our social networks. When it finds someone who is like us, it uses their information to make a recommendation to us. It’s like magic when it works. This kind of recommendation isn’t about one-time watching or one-time purchases, but about a recurring pattern.
These kinds of algorithms aren’t just mathematical. You can view them as systems, and ask yourself–how are they being implemented in your own networks? How do people find what they need (Search)? How do they get introduced to people within your network, or your conference event (Routing)? Think through each of these items and ask how, in your network, they happen–and how you can make them happen more intentionally and smoothly.

{ 0 comments… add one now }

Leave a Comment

Previous post:

Next post: