Would like to know is node.js with V8 engine suitable to be deployed on limited memory device (e.g. 256mb) and running in parallel with other process.
I read that it will hook up the resource of the machine. Is there way to limit the memory and processing usage of V8 engine itself?
256 MB is sufficient amount of RAM to run Node.js (e.g. on Linux VPS instance), assuming no other memory-hog software is run. Node has --max-stack-size argument for limiting the memory usage.
Node's single-thread evented server model generally makes efficient use of resources, but V8 due its JIT architecture is likely to use somewhat more memory than interpreted/bytecompiled implementations, such as PHP or CPython (while offering superior performance). Also, to take advantage of multiple CPU cores, multiple Node.js processes must be run (versus memory-sharing threads), effectively multiplying the memory usage, but this limitation applies to its most popular competitors as well.
In the respect of "running in parallel with other process" or "hooking up the resource of the machine", there is nothing special about running Node.js process (except the not uncommon multicore issue); it behaves similarly to any userland program. You can low-prioritize the Node.js process in OS level (e.g. with nice), but depending on your device/application, I/O can be potentially more an issue.
Purely from technical/effectiviness perspective, Erlang is probably more ideal choice for a high-level language when true multiprocessing support and high concurrency is required.