Important: This documentation covers Yarn 1 (Classic).
For Yarn 2+ docs and migration guide, see yarnpkg.com.

Package detail

@ojshua/pigato

oJshua4MIT0.0.51TypeScript support: included

An high-performance Node.js microservices framework based on ZeroMQ

pigato, nodejs, zmq, zeromq, 0mq, request-reply, amqp, mdp, micro-services

readme

PIGATO

Checkout Grenache , a DHT based high-performance microservices framework for Node.js, Ruby and Go. Supports ZeroMQ and WebSocket transports.

PIGATO - an high-performance Node.js microservices framework based on ZeroMQ

PIGATO aims to offer an high-performance, reliable, scalable and extensible service-oriented framework supporting multiple programming languages: Node.js/Io.js and Ruby.

Travis Build Status NPM version

Supported Programming Languages

  • PIGATO : PIGATO Client/Worker/Broker for Node.js / Io.js
  • PIGATO-RUBY : PIGATO Client/Worker for Ruby
  • PIGATO-GO : PIGATO Client for Go

Structure and Protocol

Actors

  • Worker : receives requests, does something and replies. A Worker offers a Service, should be a functionality as atomic as possible
  • Client : creates, pushes Requests and waits for results. A request always includes a service name and data for the Worker
  • Broker : handles Requests queueing and routing

Benefits

  • High-performance
  • Realiable, Distributed and Scalable
  • Load Balancing
  • No central point of failure
  • Multi-Worker : infinite Services and infinite Workers for each Service
  • Multi-Client : infinite Clients
  • Multi-Broker : infinite Brokers to avoid bottlenecks and improve network reliability

Features

  • Request / Reply protocol
  • Support for partial Replies
  • Client concurrent Requests
  • Client streaming Requests
  • Worker concurrent Requests
  • Worker dynamic load balancing
  • Client heartbeating for long running requests. Allows Workers to dected whenever Clients disconnect or lose interest in some request. This feature is very useful to stop long-running partial requests (i.e data streaming).

Examples

Start a Broker

node examples/broker

1) echo : simple echo request-reply

node examples/echo/worker
node examples/echo/client

2) stocks : get stocks data from yahoo

node examples/stocks/worker
node examples/stocks/client

More examples

PIGATO-EXAMPLES : a collection of multi-purpose useful examples.

Performance

PIGATO-PERF : a command-line tool to test PIGATO performances in different scenarios. i

API

Broker

PIGATO.Broker(addr, conf)

  • addr - Broker address (string, i.e: 'tcp://*:12345')
  • conf - configuration override (type=object, i.e { concurrency: 20 })
    • onStart: function to be called when the Broker start
    • onStop: function to be called when the Broker stop

Simply starts up a broker.

var Broker = require('./../index').Broker;

var broker = new Broker("tcp://*:55555");
broker.start(function() {
  console.log("Broker started");
});

Events

  • start : on Client start
  • stop : on Client stop

Worker

PIGATO.Worker(addr, serviceName, conf)

  • addr - Broker address (type=string, i.e: 'tcp://localhost:12345')
  • serviceName - service implemented by the Worker (type=string, i.e: 'echo')
    • wildcards * are supported (i.e: 'ech*')
  • conf - configuration override (type=object, i.e { concurrency: 20 })
    • prefix - sets the Worker identifier prefix
    • concurrency - sets max number of concurrent requests (type=int, -1 = no limit)
    • onConnect: function to be called when the Worker connects to the Broker
    • onDisconnnect: function to be called when the Worker disconnects from the Broker

Methods

on

Worker receives request events with 3 arguments:

  • data - data sent from the Client (type=string/object/array).
  • reply - extended writable stream (type=object)
  • opts - client request options (type=object)

reply writable stream exposes also following methods and attributes:

  • write() - sends partial data to the Client
  • end() - sends last data to the Client and completes/closes current Request
  • reject() - rejects a Request.
  • heartbeat() - forces sending heartbeat to the Broker
  • active() - returns the status of the Request (type=boolean). A Request becomes inactive when the Worker disconnects from the Broker or it has been discarded by the Client or the Client disconnects from the Broker. This is useful for long running tasks so the Worker can monitor whether or not continue processing a Request.
  • ended - tells if the Request has been ended (type=boolean).

Example

var worker = new PIGATO.Worker('tcp://localhost:12345', 'my-service');
worker.start();

worker.on('request', function(data, reply, copts) {
  for (var i = 0; i < 1000; i++) {
    reply.write('PARTIAL DATA ' + i);
  }
  reply.end('FINAL DATA');
});

// or
worker.on('request', function(data, reply, copts) {
  fs.createReadStream(data).pipe(reply);
});

Worker can change concurrency level updating its configuration. This information is carried with the heartbeat message.

Example

worker.conf.concurrency = 2;

Take note: due to the framing protocol of zmq only the data supplied to response.end(data) will be given to the client's final callback.

Events

  • start : on Worker start
  • stop : on Worker stop
  • connect : on Worker connect
  • disconnect : on Worker disconnect

Client

PIGATO.Client(addr, conf)

  • addr - Broker address (type=string, i.e: 'tcp://localhost:12345')
  • conf
    • prefix - sets the Client identifier prefix
    • autostart: automatically starts the Client (type=boolean, default=false)
    • onConnect: function to be called when the Client connects to the Broker
    • onDisconnnect: function to be called when the Client disconnects from the Broker

Methods

start

Start the Client

request

Send a Request

  • serviceName - name of the Service we wish to connect to (type=string)
  • data - data to give to the Service (type=string/object/buffer)
  • opts - options for the Request (type=object)
    • timeout: timeout in milliseconds (type=number, default=60000, -1 for infinite timeout)
    • retry: if a Worker dies before replying, the Request is automatically requeued. (type=number, values=0|1, default=0)
    • nocache: skip Broker's cache
    • workerId: ID of the Worker that must handle the Request (type=string)

Example

var client = new PIGATO.Client('tcp://localhost:12345');
client.start()

client.request('my-service', { foo: 'bar' }, { timeout: 120000 })
.on('data', function(data) {
  console.log("DATA", data);    
})
.on('end', function() {
  console.log("END");      
});

// or
client.request('my-service', 'foo', { timeout: 120000 }).pipe(process.stdout);

Clients may also make request with partial and final callbacks instead of using streams.

  • serviceName
  • data
  • partialCallback(err, data) - called whenever the request does not end but emits data
  • finalCallback(err, data) - called when the request will emit no more data
  • opts

Example

client.request('my-service', 'foo', function (err, data) {
  // frames sent prior to final frame
  console.log('PARTIAL', data);
}, function (err, data) {
  // this is the final frame sent
  console.log('FINAL', data);
}, { timeout: 30000 });

Events

  • start : on Client start
  • stop : on Client stop
  • connect : on Client connect
  • disconnect : on Client disconnect

Core Services

Core services are a set of Services that interact with a Broker via a dedicated PUB/SUB channel to extend its core functionalities.

Initialization

var broker = new PIGATO.Broker(bhost);
var csrv = new PIGATO.services.ExampleCoreService(bhost, {
  intch: broker.conf.intch // internal pub/sub channel                  
});
broker.start();
csrv.start();

Directory

PIGATO.services.Directory

Directory service ($dir) replies to Requests with the list of available Workers for a selected service.

Example

// Broadcast a message to all Workers that offer 'echo' Service
client.request(
  '$dir', 'echo', undefined, 
  function(err, workers) {
    workers.forEach(function(wid) {
      client.request('echo', 'foo', { workerId: wid });
    });
  }
);

Notes

  • when using a inproc socket the broker must become active before any queued messages.

Specification (good for RFC)

  • Worker <-> Broker heartbeating.
  • Broker tracks Worker/Client/Request relation.
  • Client MAY send heartbeat for active request. If the request is being processed by Worker, Broker forwards heartbeat to Worker.
  • Worker MAY decide to stop an inactive Request (tracks liveness for Request).
  • Client MAY assign a timeout to a Request.
  • Worker SHALL NOT send more W_REPLY (for a Request) after sending first W_REPLY message.
  • Broker SHALL force disconnect Worker if any error occurs.

Protocol

Common

  • Frame 0: Side tag (MDP.CLIENT/MDP.WORKER)
  • Frame 1: Message type (MDP.W_REQUEST, MDP.W_REPLY, MDP.W_REPLY_REJECT, ...)
  • Frame 2: Service name
  • Frame 3: Request ID (uuid)

Client request

  • Frame 4: JSON encode request data
  • Frame 5: JSON encode request options

Worker reply

  • Frame 4: Numeric status (0=OK, -1=ERROR)
  • Frame 5: JSON encode request data / error
  • Frame 6: JSON encode request options

Changelog

CHANGELOG

Roadmap

  • Add authentication support through zmq-zap ZeroMQ ZAP to trust Clients and Workers.

Follow me

Contributors

changelog

PIGATO CHANGELOG

v0.0.45

  • Minor fix

v0.0.44

  • Improved internal structures.
  • Added support for Worker liveness refresh via Client.

v0.0.43

  • Js lint
  • BREAKING: Worker and Client now must specificy 'prefix' instead of 'name'. The socket identifier is generated using 'prefix' followed by a random uuid. This fixes issues in socket identifier overlapping.
  • Support for Worker semver

v0.0.42

  • Upgraded to zmq-2.13.0 to support iojs >= 3.0
  • Removed cache support from Broker (shoulbe be refactored)
  • Broker minor refactoring
  • Switched to use an improved performance queue library

v0.0.41

  • Minor performance improvements
  • Test fixes

v0.0.40

  • Worker can see Client request options
  • Broker stop timeout (flush sockets)
  • Fix Worker call reply.end() without arguments
  • Added error emitter when Broker receives invalid messages
  • Worker heartbeating logic improvement

v0.0.39

  • Minor perf improvements
  • Load and rand broker policies
  • Wildcard fixes (maxired)
  • Symmetric behaviour for Client/Worker/Broker (maxired)
  • Add Broker onStart/onStop callbacks and start/stop events (maxired)
  • Add Worker onConnect/onDisconnect callbacks and connect/disconnect/start/stop events (maxired)
  • Add Client onConnect/onDisconnect callbacks and connect/disconnect/start/stop events (maxired)
  • More tests (maxired)
  • A bit of refactoring

v.0.0.38

  • Minor perf improvements

v.0.0.37

  • Minor changes

v.0.0.36

  • Minor changes

v.0.0.35

  • Core Services exported
  • Core Services documentation

v.0.0.34

  • Test suite refactoring for improved execution speed
  • Support for targeting a Client Request to a specific Worker using the Request option workerId
  • Minor Broker code refactoring
  • Broker dispatcher improvements
  • Refactored test directory structure
  • New test for file descriptors management
  • Fixed Broker internal request-map memory leak

v.0.0.33

  • Minor fixes
  • Stress test for file descriptors
  • Changelog moved to its own file

v.0.0.32

  • Minor fix in client.requestStream
  • Changelog and protocol specs

v.0.0.31

  • Support for opts.nocache flag in client.request : Client requests a fresh uncached reply