First of all, without knowing much about the subject you mentioned, I'm willing to bet that the premise of your question is simply wrong and that there are actually quite a lot more than 5 books on Intro to Computer Networks published by reasonably well-known publishers in the last 20 years. The world is a big place, and the publishing world is also much larger than you think, and don't forget it includes books written in languages you don't speak (yes, believe it or not, many textbooks are written in languages other than English) or that cater to an audience from a geographical region or academic subculture you have no connection with. The numbers you cited (5 books, top 15 publishers, last 20 years) are clearly just your own private speculation and have no evidence backing them, so assuming that they're wrong with high probability is one way of answering the question.
Second, another way in which your question is unfortunately not very well-formulated is that it's phrased as a multiple choice question but offers only two choices, none of which really come close to the heart of the reason why more textbooks don't get written (to the extent that there's anything that needs explaining here, since as I said in the first paragraph, a lot more textbooks actually do get written than I think you realize).
The truth is that writing textbooks is simply very hard. Academics have to decide what they will work on at any given point in their careers and what they will invest their limited time and energy in, so naturally they will opt to invest those resources in projects that will have the best impact on the world, their scientific community, and their personal careers (where these considerations are ordered differently according to what that person cares about most), and that have the best chance of success. Writing a textbook is a big project that takes anywhere between 1-4 years, and sometimes much longer (I personally know an author who took 15 years). While there are a lot of people who would like to undertake such a large project in some utopian, hypothetical sense, when it comes down to the practical task of actually sitting down and writing one, many people end up being deterred by the large effort and degree of focus that this task requires. By contrast, some alternative ways for academics to invest their time in that can be just as beneficial as writing a textbook, like working on short-to-medium term research projects, or even long-term research projects that yield intermediate deliverable products like actual publishable papers on a short-to-medium term basis, end up being a lot more suited to the psychology of many academics and hence a lot of people will prefer to focus on those other approaches to making an impact and advancing their field and career.
Finally, I have two other remarks about some of the reasons mentioned by others for why more textbooks don't get written. One is that I disagree with those who say that writing a textbook is a thankless task - my experience is that if you write a good textbook that actually helps your field in a substantial way rather than just present a slightly different point of view from 10 other textbooks, then that is likely to win you a lot of respect and admiration from colleagues, and even translate into material career benefits - promotions, speaking invitations, maybe job offers, etc. Overall, if you choose your textbook project wisely and have the talent to write a good one that makes a real contribution, it can be just as good for your career as any other thing you could be doing with an equal amount of time. It can also teach you things that doing other activities will not teach you, and make you a better and more disciplined scientist in ways that will eventually produce their own benefits far into the future. So those secondary and tertiary effects also need to be considered, but perhaps it's true that those benefits are not sufficiently appreciated by many would-be authors.
The second remark is that I also disagree with people citing the low royalties as reasons why anybody doesn't write a textbook. All the other activities that academics do when they're not writing textbooks don't pay any royalties at all, so from that point of view writing a textbook is strictly more profitable than publishing papers, writing research software or whatever. This calculation of the royalties you get divided by the number of hours it took you to write the book mentioned in some of the comments is completely nonsensical - you are paid a salary by your university while writing a textbook, just like you are paid a salary when you write your latest paper or do anything else. So, to summarize, yes the royalties are low, and certainly I don't think a sane academic will write a textbook for the royalties it will earn them, but just as certainly no one will not write a textbook because the royalties are low. If an academic decides not to write a textbook, you can be pretty sure that they made that decision for completely unrelated reasons.