High Page Faults and I/O Rate

[shotwn]

I am trying to implement a very basic RPC module using zeromq.js. It has a matching python module that works as expected when using it as python <-> python. I am trying to implement both modules on windows at the moment.
Receiving messages from python causes high amount of page faults and unexpectedly high i/o. It is especially noticeable in relatively high frequency situations. I had tried Pair and Router - Dealer couples. Both gave almost same results.

Requests: ["CALL",300,{"callback":0},"plsnogank",["report"],{"args":[false,true]}]
Answers: ["RESULT", request_id, {callback: 0}, "result"]

~10 iterations per second, 300 iterations in total.
node -> python -> node : 153.046 page faults, 8MB/s avg total I/O during iterations.
python -> python -> python: 6.600 page faults, 20kB/s avg total I/O during iterations.
node -> python: 55.289 page faults, 1.88kB/s avg total I/O during iterations.

~100 iterations per second, 300 iterations in total.
node -> python -> node : 151.445 page faults, 65MB/s avg total I/O during iterations.
python -> python -> python: 6.070 page faults, 190kB/s avg total I/O during iterations.
node -> python : 57.610 page faults, 19kB/s avg total I/O during iterations.

~100 iterations per second, 3000 iterations in total.
node -> python -> node : 335.351 page faults, 62MB/s avg total I/O during iterations.
python -> python -> python: 6.126 page faults, 190kB/s avg total I/O during iterations.
node -> python : 27.294 page faults, 22kB/s avg total I/O with 18kB/s to 800kB/s spikes.

I did my observations through Process Hacker and Windows Performance Monitor. When I don't send anything from python to node everything gets relatively calmer.
I am pretty new in everything zmq but I/O rate, marked as "other bytes" is extremely high compared to everything else working in my system.

What is causing this ? Is it a bug or is it about how I implement zmq to my code ?

Here is my code.
bridge.js
test.bridge.js

It is maybe not very relevant but python part.
bridge.py
test.rpc.py

[rolftimmermans]

Recently we have released 6.0 beta. It features a new API that addresses some fundamental issues with the previous API and also addresses a number of stability bugs. To make upgrading easier it includes a compatibility layer with versions 4.x/5.x. It would be great if you could give the latest version a spin to see if this solves this particular issue. If you run into any problems with it, feel free to report it here or in a new issue.