(Arun Murugappan) #1


hy is the graph as shown below in the attachment. Explain

(Mark Walkom) #2

Which graph?

(Arun Murugappan) #3

the graph in the attachment. Cost vs users

(Christian Dahlqvist) #4

I think you need to elaborate a bit more on your question as it is not clear what you are asking. It also seems quite general and not directly Elasticsearch related. Maybe this link is useful?

(Arun Murugappan) #5

It is related to the link that you have attached. Thanks. It is related to ES. As ES scales well as compared to rdbms. My question there in lies, in horizontal scaling--why does the system cost rise at a constant rate but in vertical scaling the system cost rise steeply. Likeise application performance is maintained in horizontal scaling but dips a little in vertical scaling. Why so?

(Christian Dahlqvist) #6

Typically scaling out by buying cheap commodity hardware is a cheaper way to add computing resources rather than buying larger and larger servers, which tend to get more expensive. There is also often a limit to the size of server you can get, so when your needs exceed what one maxed out server can handle you might find yourself in a tricky situation. Why the graph dips at the end I do not know as I did not create the graph.

(Arun Murugappan) #7

Thanks Christian. With respect to cost, probably yes. If there is limit to the size of servers (vscal), isnt there a limit to the number of servers as well(hscal)?

(Christian Dahlqvist) #8

There may naturally be limits to horizontal scalability as well, but it often depends on what type of system you are running.

(Arun Murugappan) #9

okay. yes. But there is always preference to horizontal scaling than vertical scaling. Am I right?

(Christian Dahlqvist) #10

I do not think there is any easy answer to that as it depends on the system you are deploying.

(Arun Murugappan) #11

got that

(system) #12

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.