Tell me more ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

People conencts to a Bank using SSL.

Bank is openning its 443 port.

but there is only 1 port of 443 , but it is a server so I guess he is using also the other ports.

But thats means that it can only have ~65k connections at same time.

Or , does it actually manage all the ssl requests on the same port(443) but opens new thread for each request.

But thats means that every customer will notice a big delay since there are other users which consume info .

I must be wrong here.

Is this how things done ?

share|improve this question
And that's why you have more than one server. – Nifle Dec 26 '12 at 9:23
@Nifle so whic part is the correct , the first one or the second ? ( in my question) – Royi Namir Dec 26 '12 at 9:36
Everybody uses the same port. And it your users experience delay because the webserver(s) can't keep up you add more servers. – Nifle Dec 26 '12 at 11:56
@Nifle Im not a system guy , just asking. – Royi Namir Dec 26 '12 at 12:01
No harm in asking. – Nifle Dec 26 '12 at 12:04

closed as not constructive by Nifle, TFM, techie007, Simon Sheehan, CharlieRB Dec 26 '12 at 18:08

As it currently stands, this question is not a good fit for our Q&A format. We expect answers to be supported by facts, references, or specific expertise, but this question will likely solicit debate, arguments, polling, or extended discussion. If you feel that this question can be improved and possibly reopened, see the FAQ for guidance.

1 Answer

up vote 2 down vote accepted

There can be an infinite number of incoming connections to a single port (subject to OS implementation limits) as long as the other end can be identified with unique IP-port combinations. The server do not need to open a unique local port for each client. So the first bit if your second part is right.

Also, most operating systems can handle tens of thousands of threads without much perceived performance degredation. As long as the request handlers are coded with the usual concurrent programming requirements, the threads should be almost independent from each other. I guess most banking application will synchronise on a database as opposed to much in-memory objects, so the chance of a thread being blocked waiting for another is even lower.

share|improve this answer

Not the answer you're looking for? Browse other questions tagged or ask your own question.