Do We Need Math in Networking?

I should have known better, but I couldn’t resist being pulled into a Twitter spat around the question “whether networking engineers need to know something about math” a long while ago.

Before going into the details, let’s start with Wikipedia definition: “Engineering is the use of scientific principles to design and build machines, structures, and other things” including “specific emphasis on particular areas of applied mathematics, applied science, and types of application”.

So feel free to believe that you don’t need any math or other science (because there’s very little science behind what we do in networking) in your job, in which case you might want to stop reading… but then at least please think twice about your job title.

Still here? Let me give you just a few examples of where we could use math:

  • Replacing common sense and gut feeling with verifiable facts. I was told that "having two links does not result in 100% uptime” is plain common sense. Unfortunately, this particular type of common sense seems to be rare. How about using some basic reliability theory? It’s not hard, and it helps you calculate what you can expect based on the structure of your network.
  • Debunk vendor claims. If you know what can be done, you’ll quickly spot vendors making impossible claims.
  • Optimizing your network. When I was talking with Cariden engineers a long while ago, they told me how large service providers saved millions by shifting traffic across expensive international links by optimizing routing metrics. Of course you could take a million-of-monkeys approach to the problem, or figure out the math behind it… or buy a magic black box with a “Trust Me” label on it from your favorite vendor.
  • Understanding queues. Queuing theory was admittedly the hardest topic Rachel Traylor addressed in her “math-in-networking” series… but maybe it’s time we figure out how things work instead of relying on vendor best practices (which often depend on what box they’re trying to sell you this week).

Of course you don’t need math (or understanding the principles) in every network. After all, you don’t need structural analysis to set up a camping tent… but a good engineer should know when a problem he’s solving is simple enough to wing it, and when he should pull out his toolbox. Claiming that you don’t need a toolbox because you don’t know what’s in it doesn’t make much sense.

5 comments:

  1. Math? Do wee need more than the most important formula:n*(n-1)/2?

    Queuing/BW? Should we calculate something or simply throw in more BW?

    Logic. Yes, sure. The most important part of the math here.

    I would say math is important just to eliminate those who are not smart enough...

    Joking.
  2. I have been looking to the answer to this question fir over18 yrs since started career in networking.
    Math is really hidden in all networking protocols. Even more most CCIEs that have seem to know very little about the science behind it on not to mention the math. Great topic as usual and first of a kind that i see trying to bring some science into networking
  3. I'm surprised nobody said binary arithmetic, I'm sure that most of you had issues explaining how to create a proper subnet address planning , once people moves out of a /24 you start to see lot of funny things :-)
    Replies
    1. This is expert level that's why I mentioned n*(n-1)/2 ;)
  4. I'd use the cliche to answer the question: It depends.

    If one is working with more practical staffs, it's rare to use any math in projects. calc man-day would be one of the hardest things in math.

    But if they're things related to protocols and why they're better or worse than counterparts, it's impossible to understand them without math. e.g. nth-shortest path, queuing, and etc.
Add comment
Sidebar