Mime Parsing performance

[rajat shukla]

Hi,

I have developed a Mime parsing code. It takes around 46ms to parse a 200kb message. Is it up to the standards?

[Ulf Dittmer]

Which standards?

[Jeanne Boyarsky]

46ms sounds like nothing - unless you are doing it a lot of times of course. Do you have any performance requirements?

[rajat shukla]

I am using apache mime parsing API. A message queue is pre populated with 2000 msg each of size 200KB. It is a single threaded Environment. The parser takes one message at a time and parses it. The client requirements is that each message should take max of 20ms. but I could only bring the numbers to 46ms. Is it a valid requirement or the client is pushing too hard. Or is there any other API which could give me better performance with the current hardware I am having.(Intel dual core with 2 GB Ram)

[Ulf Dittmer]

"2 GB of RAM" sounds like a desktop machine - maybe the server will be faster? Not sure if 20ms is reasonable; what is the actual requirement behind it that drives this number?

[William Brogden]

What happens to the parsed data? Surely that has potential for slowing things down.

Are you parsing as a character stream or a byte stream? Character streams of course involve a time consuming unicode conversion.

Bill

[Jeanne Boyarsky]

Also, the server is likely to not be single threaded which cuts the time down further.