# What kind of distribution is that?

Could you tell me what kind of distribution is that? I can also provide the original data if needed.

-
Did you try a biased normal distribution? –  Jon Feb 26 '12 at 16:44
Quite triangular. –  Did Feb 26 '12 at 16:46

As Didier says, it looks like a triangular distribution -- specifically, if we ignore the long tail, it could be the sum of two independent variables distributed uniformly between 0 and about 256, plus perhaps some lower-order noise.

It could also be the sum of a random number of uniformly distributed $(0,256)$ variables, which would explain both the tail and the slight sudden drop in frequency that appears to happen right at the apex. It's not easy to estimate the distribution of the number-of-bytes from the graph, except that $2$ clearly dominates, and $1$ is more likely than $>2$.

Whether or not this is a sensible hypothesis in your case depends entirely on the process that created the your data.

-
+1, if only for the last paragraph. –  Did Feb 26 '12 at 17:53