We're not saying if it's typical or not typical, we're saying exactly that, that it all depends A LOT on your hardware and architecture configuration. Such environments may not be typical for a single user who is just doing some basic stuff, or for a small company or others, but it is very typical for any decent company who has high traffic volume and doesn't risk their business with low end hardware.
As we all know in the list, there are MANY Telco's out there which use Kannel on their infrastrucures, among them Telefonica which is one of the biggest of the world. I have never seen anyone on the list mentioning he works on any carrier, and most are probably not allowed to do so and say "hey, we use Kannel", but we all know lots of carriers use Kannel at least on parts of their infrastructure. And I can't imagine those big carriers using just a low end server... Maybe the biggest Kannel user base corresponds to those carriers, just that they don't participate on the list, or they just don't mention it... so if that's the case, then the typical Kannel usage is not the one we mostly see on the list... This is all possibilities, of course, but you get the point. What started the discussion is not if the numbers you gave are the ones the "normal user" may encounter, which maybe it's possible. What we're saying is that it doesn't mean that they are the max or whatever numbers that Kannel can handle, which is what someone could think from you original statement. On X environment, with X server and X tests, it gave X results. That's fine. But saying that results will always be similar, or that the proportion of MOs it can handle will be X% of the MTs, etc, etc, is not reasonable. That's on _your_ case, with _your_ server configurations, and with _your_ Kannel configurations. Other can vary a lot. Regards, Juan 2010/8/14 Nikos Balkanas <[email protected]>: > Again, I just replied to a user asking about typical conditions. Some of the > exotic things that Alex mentioned, like RAID or solid-state fs or clustered > DBs are not considered typical, and are hardly what the user was asking for > (else he would have mentioned them). Nevertrheless, it would be interesting > to get some figures on these setups too, so that people can now the real > benefit before spending a lot of money on them. > > Disagreeing is all democratic. Doing it responsibly and not being a > nillihist is a matter of choice.
