[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
Re: [jetty-dev] NIO inefficiency?
|
On Monday, August 12, 2013 at 9:29 AM, Viktor Szathmary wrote:
The workload is lots of small GET requests from browser all over the internet (a bunch of params and cookies to parse), the response is a few kb of JSON dynamically generated and gzipped. The servers (EC2 m1.large instances, sitting behind an Elastic Load Balancer) get around 5-600 QPS each. We spend about half a millisecond on average in our application code to generate the response. The rest of our latency is in the stack (jetty/linux/virtualization/network) – a total of around 1.5ms per request as measured by the ELB.
To clarify: the ELB uses HTTP/1.1 connections, keeps them open and reuses them to serve multiple requests from client browsers. Thus in general we only have a moderate number of connections open (e.g. ~40-60). Not sure if ELB does any pipelining though (I would think not)…
Regards,
Viktor
