microgen-cote
Advanced tools
microgen-cote lets you write zero-configuration microservices in Node.js without nginx, haproxy, redis, rabbitmq or anything else. It is batteries — and chargers! — included.
Join us on
for anything related to microgen-cote.
in time-service.js...
const microgen-cote = require('microgen-cote');
const timeService = new microgen-cote.Responder({ name: 'Time Service' });
timeService.on('time', (req, cb) => {
cb(new Date());
});
in client.js...
const microgen-cote = require('microgen-cote');
const client = new microgen-cote.Requester({ name: 'Client' });
client.send({ type: 'time' }, (time) => {
console.log(time);
});
You can run these files anyway you like — on a single machine or scaled out to hundreds of machines in different datacenters — and they will just work. No configuration, no third party components, no nginx, no kafka, no consul and only Node.js. microgen-cote is batteries — and chargers — included!
Make sure to check out the e-commerce case study that implements a complete e-commerce application with microservices using microgen-cote. It features;
microgen-cote plays very well with Docker, taking advantage of its network overlay features. The case study implements a scalable microservices application via Docker and can scale to multiple machines.
Tomorrow belongs to distributed software microservices.
As CPU performance is heavily dictated by the number of cores and the power of
each core is already at its limits, distributed computing will decide how your
application performs. Distributed systems Microservices also pose great
architectural benefits such as fault-tolerance and scalability.
Components of such a distributed system microservices should be able to
find other components
zeroconf and communicate over a set of conventions. Sometimes they may work
as a cluster, may include a pub/sub mechanism, or a request/response mechanism.
microgen-cote brings you all the advantages of distributed software microservices.
Think of it like homing pigeons.
microgen-cote allows you to implement hassle-free microservices by utilizing auto-discovery and other techniques. Typically, in a microservices system, the application is broken into smaller chunks that communicate with each other. microgen-cote helps you build such a system by providing you several key components which you can use for service communication.
In a way, microgen-cote is the glue that's most necessary between different microservices. It replaces queue protocols and service registry software by clever use of IP broadcast/IP multicast systems. It's like your computer discovering there's an Apple TV nearby. This means, microgen-cote needs an environment that allows the use of IP broadcast or multicast, in order to scale beyond a single machine. Most bare-metal systems are designed this way, however, cloud infrastructure like AWS needs special care, either an overlay network like Weave, or better yet, just, Docker — which is fortunately the way run all of our software today anyway. That's why Docker is especially important for microgen-cote, as it enables microgen-cote to work its magic.
microgen-cote also replaces HTTP communication. Microservices architecture is meant for hundreds of internal services communicating with each other. That being the case, a protocol like HTTP is cumbersome and heavy for communication that doesn't need 90% of HTTP's features. Therefore, microgen-cote uses a very light protocol over plain old TCP sockets for communication, making it fast, effective and most importantly, cheap.
microgen-cote is a Node.js library for building microservices applications. It's available as an npm package.
Install microgen-cote locally via npm:
npm install microgen-cote
Whether you want to integrate microgen-cote with an existing web application — e.g. based on express.js as exemplified here — or you want to rewrite a portion of your monolith, or you want to rewrite a few microservices with microgen-cote, all you need to do is to instantiate a few of microgen-cote's components (e.g. Responder, Requester, Publisher, Subscriber) depending on your needs, and they will start communicating automatically. While one component per process might be enough for simple applications or for tiny microservices, a complex application would require close communication and collaboration of multiple microservices. Hence, you may instantiate multiple components in a single process / service / application.
The most common scenario for applications is the request-response cycle. Typically, one microservice would request a task to be carried out or make a query to another microservice, and get a response in return. Let's implement such a solution with microgen-cote.
First, require microgen-cote;
const microgen-cote = require('microgen-cote');
Then, instantiate any component you want. Let's start with a Requester that
shall ask for, say, currency conversions. Requester and all other components
are classes on the main microgen-cote object, so we instantiate them with the new
keyword.
const requester = new microgen-cote.Requester({ name: 'currency conversion requester' });
All microgen-cote components require an object as the first argument, which should at
least have a name property to identify the component. The name is used mainly
as an identifier in monitoring components, and it's helpful when you read the
logs later on as each component, by default, logs the name of the other
components they discover.
Requesters send requests to the ecosystem, and are expected to be used
alongside Responders to fulfill those requests. If there are no Responders
around, a Requester will just queue the request until one is available. If
there are multiple Responders, a Requester will use them in a round-robin
fashion, load-balancing among them.
Let's create and send a convert request, to ask for conversion from USD into
EUR.
const request = { type: 'convert', from: 'usd', to: 'eur', amount: 100 };
requester.send(request, (err, res) => {
console.log(res);
});
You can save this file as client.js and run it via node client.js.
client.js file.
const microgen-cote = require('microgen-cote');
const requester = new microgen-cote.Requester({ name: 'currency conversion requester'});
const request = { type: 'convert', from: 'usd', to: 'eur', amount: 100 };
requester.send(request, (err, res) => {
console.log(res);
});
Now this request will do nothing, and there won't be any logs in the console, because there are no components to fulfill this request and produce a response.
Keep this process running, and let's create a Responder to respond to currency conversion requests.
We first instantiate a Responder with the new keyword.
const responder = new microgen-cote.Responder({ name: 'currency conversion responder' });
As detailed in Responder, each Responder is also an instance of
EventEmitter2. Responding to a certain request, let's say convert, is the
same as listening to the convert event, and handling it with a function that
takes two parameters: a request and a callback. The request parameter holds
information about a single request, and it's basically the same request object
the requester above sent. The second parameter, the callback, expects to be
called with the actual response.
Here's how a simple implementation might look like.
const rates = { usd_eur: 0.91, eur_usd: 1.10 };
responder.on('convert', (req, cb) => {
cb(null, req.amount * rates[`${req.from}_${req.to}`]);
});
Now you can save this file as conversion-service.js and run it via
node conversion-service.js on a separate terminal.
conversion-service.js file.
const microgen-cote = require('microgen-cote');
const responder = new microgen-cote.Responder({ name: 'currency conversion responder' });
const rates = { usd_eur: 0.91, eur_usd: 1.10 };
responder.on('convert', (req, cb) => {
cb(null, req.amount * rates[`${req.from}_${req.to}`]);
});
As you run the service, you will immediately see the first request in
client.js being fulfilled and logged to the console. Now you can take this
idea and build your services on it.
Notice how we didn't have to configure IP addresses, ports, hostnames, or anything else.
Note: By default, every
Requesterwill connect to everyResponderit discovers, regardless of the request type. This means, everyRespondershould respond to the exact same set of requests, becauseRequesters will load-balance requests between all connectedResponders regardless of their capabilities, i.e, whether or not they can handle a given request.
If you have multiple Responders with varying response handlers, you will
experience lost requests. In microgen-cote, this separation between responsibilities is
called segmentation, or partitioning. If you wish to segment your requests in
groups, you can use keys. Check out keys for a detailed guide on how
and when to use segmentation.
One of the benefits of a microservices approach is its ease of use as a tool for tasks that previously required serious infrastructural investments. Such a task is managing updates and tracking changes in a system. Previously, this required at least a queue infrastructure with fanout, and scaling and managing this technological dependency would be a hurdle on its own.
Fortunately, microgen-cote solves this problem in a very intuitive and almost magical way.
Say, we need an arbitration service in our application which decides currency rates, and whenever there's a change within the system, it should notify all the instances of conversion services, so that they facilitate the new values.
Of course, the arbitration service would be API driven, and would receive the
new rates over another request so that for example an admin can enter the values
through a back office application. The arbitration service should take this
update and basically forward it to every conversion service. In order to achieve
this, the arbitration service should have two components: one Responder for
the API updates and one Publisher for notifying the conversion services. In
addition to this, the conversion services should be updated to include a
Subscriber. Let's see this in action.
A simple implementation of such a service would look like the following. First,
we require microgen-cote and instantiate a responder for the API. Since we now have two
responders, arbitration API and currency conversion responder, we need to
introduce service segmentation by using key property. If we had no keys in
our examples, some requests from our client.js would end up in
currency conversion responder and we would get a correct response, but some
other requests would end up in arbitration API, and since arbitration
responder isn't listening to 'convert' events, the request would remain
unanswered.
arbitration-service.js
const microgen-cote = require('microgen-cote');
const responder = new microgen-cote.Responder({ name: 'arbitration API', key: 'arbitration' });
Let's say we keep the rates in a local variable. This could just as well be a database call, but for the sake of simplicity let's keep this local.
const rates = {};
Now the responder shall respond to an rate updated request, allowing admins to
update it from a back office application. The backoffice integration isn't
important at this moment, but here is an example how back offices could
interact with microgen-cote responders in the backend.
Basically, this service should have a responder to take in the new rates for a
currency exchange.
responder.on('update rate', (req, cb) => {
rates[req.currencies] = req.rate; // { currencies: 'usd_eur', rate: 0.91 }
cb(null, `changed ${req.currencies} rate to ${req.rate}`);
});
We now have the rates, but the rest of the system, namely, the conversion
services aren't aware of this change yet. In order to update them of the
changes, we should create a Publisher.
const publisher = new microgen-cote.Publisher({ name: 'arbitration publisher' });
Now whenever there's a new rate, we should utilize this Publisher. The
update rate handler thus becomes:
responder.on('update rate', (req, cb) => {
rates[req.currencies] = req.rate;
cb(null, `changed ${req.currencies} rate to ${req.rate}`);
publisher.publish('rate updated', req);
});
arbitration-service.js file.
const microgen-cote = require('microgen-cote');
const responder = new microgen-cote.Responder({ name: 'arbitration API', key:'arbitration' });
const publisher = new microgen-cote.Publisher({ name: 'arbitration publisher' });
const rates = {};
responder.on('update rate', (req, cb) => {
rates[req.currencies] = req.rate;
cb(null, `changed ${req.currencies} rate to ${req.rate}`);
publisher.publish('rate updated', req);
});
Since currently there are no subscribers in this system, nobody will be notified
of these changes. In order to facilitate this update mechanism, we need to go
back to our conversion-service.js and add a Subscriber to it.
A Subscriber is a regular microgen-cote component, so we instantiate it with the
following:
const subscriber = new microgen-cote.Subscriber({ name: 'arbitration subscriber' });
Put this line in conversion-service.js.
Subscriber also extends EventEmitter2, and although these services might run
in machines that are continents apart, any published updates will end up in a
Subscriber as an event for us to consume.
Here's how we might update conversion-service.js to listen to updates from the
arbitration service.
subscriber.on('rate updated', (update) => {
rates[update.currencies] = update.rate;
});
Let's not forget to change the use of requester and responder in our conversion service and in our client to use segmentation key.
conversion-service.js
const responder = new microgen-cote.Responder({ name: 'currency conversion responder', key: 'conversion' });
client.js
const requester = new microgen-cote.Requester({ name: 'currency conversion requester', key: 'conversion' });
That's it! From now on, this conversion service will synchronize with the arbitration service and receive its updates. The new conversion requests after an update will be done over the new rate.
conversion-service.js file.
const microgen-cote = require('microgen-cote');
const responder = new microgen-cote.Responder({ name: 'currency conversion responder', key: 'conversion' });
const subscriber = new microgen-cote.Subscriber({ name: 'arbitration subscriber' });
const rates = { usd_eur: 0.91, eur_usd: 1.10 };
subscriber.on('rate updated', (update) => {
rates[update.currencies] = update.rate;
});
responder.on('convert', (req, cb) => {
const convertedRate = req.amount * rates[`${req.from}_${req.to}`];
cb(null, `${req.amount} ${req.from} => ${convertedRate} ${req.to}`);
});
client.js file.
const microgen-cote = require('microgen-cote');
const requester = new microgen-cote.Requester({ name: 'currency conversion requester', key:'conversion' });
const request = { type: 'convert', from: 'usd', to: 'eur', amount: 100 };
requester.send(request, (err, res) => {
console.log(res);
});
microgen-cote hosts a number of components that together let you implement microservice communication. Below, you will find several examples on how to make use of each component.
By default, every component can discover and interact with every other
component. This may not be desirable under certain conditions whereas security
and network performance is of importance, so one can segregate or partition
component clusters with keys and environments provided in configuration
objects.
Also, all components support namespaces. Given as a property of the
configuration object to the constructor, components adhere and act on namespaces
if provided, and ignore other messages. Namespaces are also handy in that they
let you wire a namespaced socket.io connection to the front-end. In other words,
the namespaces here also serve as socket.io namespaces.
Requester queues requests until a Responder is available, and once so, it delivers the request. Requests will be dispatched to Responders in a round-robin way.
Example:
const microgen-cote = require('microgen-cote');
const randomRequester = new microgen-cote.Requester({
name: 'Random Requester',
// namespace: 'rnd',
// key: 'a certain key',
requests: ['randomRequest'],
});
setInterval(() => {
const req = {
type: 'randomRequest',
val: Math.floor(Math.random() * 10),
};
randomRequester.send(req, (res) => {
console.log('request', req, 'answer', res);
});
}, 5000);
Requesters also support Promises, which gives you great flexibility when
working with promise-based libraries or when you want to chain multiple
Requesters and Responders.
Example with promises:
const microgen-cote = require('microgen-cote');
const randomRequester = new microgen-cote.Requester({ name: 'Random Requester' });
const makeRequest = (req) => randomRequester.send(req);
const req = {
type: 'randomRequest',
val: Math.floor(Math.random() * 10),
};
makeRequest(req)
.then(console.log)
.catch(console.log)
.then(process.exit);
Example with async / await:
const microgen-cote = require('microgen-cote');
const randomRequester = new microgen-cote.Requester({ name: 'Random Requester' });
async function makeRequest () {
const req = {
type: 'randomRequest',
val: Math.floor(Math.random() * 10),
};
const response = await randomRequester.send(req);
console.log(response);
process.exit();
}
makeRequest();
A timeout could be configured for all Requesters as an environment variable
COTE_REQUEST_TIMEOUT, or in advertisement options for specific Requester,
or in a property called __timeout in first argument of requester.send
method. Latter setting overrides former. Timeout is specified in milliseconds.
As environment variable for all requesters:
COTE_REQUEST_TIMEOUT=1000 node service.js
In advertisement settings:
new microgen-cote.Requester({ name: `Requester with timeout`, timeout: 1000 });
In send data:
requester.send({ type: 'find', __timeout: 2000 });
Responder is a component for responding to certain requests from a Requester.
It's a descendant of EventEmitter2, and requests are regular events, therefore
may be wildcarded or namespaced.
Responder may be used to add new modules to existing web servers /
applications without ever changing the main server code. Only a Requester
will be able to utilize a Responder.
You can use a Responder with a Sockend component to open a flexible API
channel for the front-end. This greatly reduces time-to-market by providing a
direct API for your front-end applications.
Example:
const microgen-cote = require('microgen-cote');
// Instantiate a new Responder component.
const randomResponder = new microgen-cote.Responder({
name: 'Random Responder',
// namespace: 'rnd',
// key: 'a certain key',
respondsTo: ['randomRequest'], // types of requests this responder
// can respond to.
});
// request handlers are like any event handler.
randomResponder.on('randomRequest', (req, cb) => {
const answer = Math.floor(Math.random() * 10);
console.log('request', req.val, 'answering with', answer);
cb(null, answer);
});
Responders also support Promises, , which gives you great flexibility when
working with promise-based libraries or when you want to chain multiple
Requesters and Responders.
Example with promises:
responder.js
const microgen-cote = require('microgen-cote');
const UserModel = require('UserModel'); // a promise-based model API such as
// mongoose.
const userResponder = new microgen-cote.Responder({ name: 'User Responder' });
userResponder.on('find', (req) => UserModel.findOne(req.query));
requester.js
const microgen-cote = require('microgen-cote');
const userRequester = new microgen-cote.Requester({ name: 'User Requester' });
userRequester
.send({ type: 'find', query: { username: 'foo' } })
.then((user) => console.log(user))
.then(process.exit);
Example with async / await
responder.js
const microgen-cote = require('microgen-cote');
const UserModel = require('UserModel'); // a promise-based model API such as
// mongoose.
const userResponder = new microgen-cote.Responder({ name: 'User Responder' });
userResponder.on('find', (req) => UserModel.findOne(req.query));
requester.js
const microgen-cote = require('microgen-cote');
const userRequester = new microgen-cote.Requester({ name: 'User Requester' });
async function makeRequest() {
const user = await userRequester.send({ type: 'find', query: { username: 'foo' });
console.log(user);
process.exit();
}
makeRequest();
Publisher is a component for publishing certain events with arbitrary data.
It may be used as a distributed EventEmitter. It may also be used in a
scenario where some components need to be notified of updates, such as new
tweets, etc. instead of polling for them. Only a Subscriber will get
notifications from a Publisher.
The messages Publishers publish are volatile in that if there are no
Subscribers listening, they are lost.
Publishers may be used in conjunction with a Sockend component, in which
case the front-end clients will be notified of the events published. This is a
very cool real-time communication mechanism for your apps with no proprietary
technology like Meteor.
Example:
const microgen-cote = require('microgen-cote');
// Instantiate a new Publisher component.
const randomPublisher = new microgen-cote.Publisher({
name: 'Random Publisher',
// namespace: 'rnd',
// key: 'a certain key',
broadcasts: ['randomUpdate'],
});
// Wait for the publisher to find an open port and listen on it.
setInterval(function() {
const val = {
val: Math.floor(Math.random() * 1000),
};
console.log('emitting', val);
// publish an event with arbitrary data at any time
randomPublisher.publish('randomUpdate', val);
}, 3000);
Subscriber subscribes to events emitted from a Publisher.
Example:
const microgen-cote = require('microgen-cote');
const randomSubscriber = new microgen-cote.Subscriber({
name: 'Random Subscriber',
// namespace: 'rnd',
// key: 'a certain key',
subscribesTo: ['randomUpdate'],
});
randomSubscriber.on('randomUpdate', (req) => {
console.log('notified of ', req);
});
Sockend is the glue for carrying all the possibilities of microgen-cote to the next
level with WebSockets over socket.io. Sockend makes Responders and
Publishers available to the front-end and adhere to socket.io namespaces.
It's the magic and the lost link for microservices. Without any configuration,
you can expose APIs directly to the front-end.
Example:
index.html
<script src="/socket.io/socket.io.js"></script>
<script>
let socket = io.connect();
let socketNamespaced = io.connect('/rnd');
socket.on('randomUpdate', function(data) {
console.log(data);
});
setInterval(function() {
let req = {
val: Math.floor(Math.random() * 10),
};
let req2 = {
val: Math.floor(Math.random() * 10),
};
let req3 = {
val: Math.floor(Math.random() * 10),
};
let req4 = {
val: Math.floor(Math.random() * 10),
};
socket.emit('randomRequest', req, function(data) {
console.log('normal', req.val, data);
});
socketNamespaced.emit('randomRequest', req2, function(data) {
console.log('ns', req2.val, data);
});
socket.emit('promised request', req3, function(err, data) {
console.log('normal promised', req3.val, err, data);
});
socketNamespaced.emit('promised request', req4, function(err, data) {
console.log('ns promised', req4.val, err, data);
});
}, 3000);
</script>
sockend.js
const microgen-cote = require('microgen-cote'),
app = require('http').createServer(handler),
io = require('socket.io').listen(app),
fs = require('fs');
io.on('connection', (socket) => {
socket.join('room1');
});
app.listen(process.argv[2] || 5555);
function handler(req, res) {
fs.readFile(__dirname + '/index.html', (err, data) => {
if (err) {
res.writeHead(500);
return res.end('Error loading index.html');
}
res.writeHead(200);
res.end(data);
});
};
const sockend = new microgen-cote.Sockend(io, {
name: 'Sockend',
// key: 'a certain key'
});
To connect responder and sockend, you need to add respondsTo: parameter to your options.
const randomResponder = new Responder({
name: 'randomRep',
respondsTo: ['randomRequest', 'promised request'], // types of requests this responder
// can respond to.
});
To connect publisher and sockend, you need to add broadcasts: parameter to your options
let randomPublisher = new Publisher({
name: 'randomPub',
broadcasts: ['update1', 'update2'],
});
If your socket is connected to a namespace, you need to add the same namespace to the components that you want to expose.
Even though socket.io are prefixed with /, with sockend you need to omit /.
socket.io-client
io.connect('/rnd'); // namespace in socket.io is declared as '/rnd'
sockend component
const randomResponder = new Responder({
name: 'randomRep',
namespace: 'rnd', // with sockend, we we omit the '/' and use just 'rnd'
respondsTo: ['randomRequest', 'promised request'], // types of requests this responder
// can respond to.
});
You can check complete code for Responders and Publishers in the examples folder. Fire them up
on default or 'rnd' namespace and watch them glow with magic on
http://localhost:5555.
If you want to see more complete example of microservices with sockend integration, check out the e-commerce case study
Sockend supports socket.io rooms. All you need to do is add a __rooms or __room attribute to
the published message.
const randomPublisher = new microgen-cote.Publisher({
name: 'Random Publisher',
// namespace: 'rnd',
// key: 'a certain key',
broadcasts: ['randomUpdate'],
});
randomPublisher.publish('randomUpdate', { val: 500, __rooms: ['room1', 'room2'] });
randomPublisher.publish('randomUpdate', { val: 500, __room: 'room1' });
Monitor is the "top" of microgen-cote. It lists all the daemons it discovers regardless
of namespace or key. Run examples/monitor.js and see all your active microgen-cote
daemons.
microgen-cote also has an infant of a monitoring tool that displays the microgen-cote ecosystem
running in your environment in a nice graph. Run examples/monitoring-tool.js
and navigate to http://localhost:5555 in your browser to see your microgen-cote network
graph in action.
While microgen-cote is extremely simple to get started, the requirements for a system running in production may demand further tweaking and advanced settings. Here are some of the advanced features of microgen-cote, which can be adjusted on several levels — as environment variables, as direct settings for the microgen-cote module when requiring it, or as direct settings for each component.
Until now, we only saw instantiating microgen-cote components with a single argument. In fact, all microgen-cote components have two constructor parameters. The first is used as the advertisement configuration which controls the data being advertised for auto-discovery. The second parameter is the discovery configuration and it controls the network-layer configuration and environments for components.
We'll see more details in the following section.
microgen-cote works over IP broadcast or multicast. This means, for example, in an office network where there are several developers running a project based on microgen-cote in their local machines, there might be chaos. Components on a developer's machine will discover other components on another developer's machine. This probably is not a desired effect, and fortunately microgen-cote offers a way to combat this.
By passing in an environment property to the configuration object of a
component, one can control the scope of auto-discovery for that particular
component. Components that are not in the same environment will ignore each
other. This effectively creates network partitions.
environments can be set as an environment variable COTE_ENV.
Running a service with
COTE_ENV=developer-1 node service.js
sets all the components within that service to use developer-1 as an
environment. This makes sure that however many modules service.js makes use
of, they all will share the same environment, so this is the safest way to
specify environments.
The other way to specify an environment is using the configuration argument
to microgen-cote, given when requiring microgen-cote in the first place. Since Node.js modules are
read and executed once from the disk, you need to make sure to pass in the configuration at least once, during the first require call. The subsequent
requires to microgen-cote will return the same module, which already has your
configuration. If you have a bootstrap in your application that runs as the
first thing in the application, it might be a good idea to put this config
there.
const microgen-cote = require('microgen-cote')({ environment: 'developer-2' });
Now the components in these services won't discover and communicate with each other.
Another place this comes handy is multiple environments running on a single
machine. Say you have a machine for your QA needs, where you host several
environments for different tests, e.g. integration and qa. Again, components
from different environments would mix up. Using a parametric environment, in
this case, solves this problem.
microgen-cote has another mechanism to create partitions called keys.
Since every component discovers and tries to communicate with every other
component on the horizon (this is called a "mesh network"), if different
services request and respond to different types of messages, you will
experience lost messages. In other words, if service A responds to messages X
and Y, and service B responds to messages Z and T, you will lose half of the
messages, because messages Z and T will also end up at service A, but it won't
know how to handle them. The same is true for service B: messages X and Y
will end up at it, but service B won't know how to respond to them.
Keys are useful in this scenario: requesters and responders around service A and messages X and Y should use one particular key, and requesters and responders around service B and messages Z and T should use another key. In this case, no messages will be lost, and the services will be segregated.
In our experience, the best way to segregate services is to follow the
principles of domain-driven design. In this regard, for example, each domain
could have its own key. If you need more granular services, you should use
multiple keys inside the same domain. The principle is to ensure distinct keys
for a distinct set of messages. In other words, keys should represent a
distinct set of requests.
Please refer to [Creating the arbitration service] (https://github.com/dashersw/microgen-cote#creating-the-arbitration-service) for an example of keys in action.
keys are given as parameters to the configuration objects.
When deciding whether to create a connection to another service, microgen-cote components
make use of keys and environments together. Therefore, two components with
exact same environments with different keys wouldn't be able to communicate.
Think of it as ${environment}_${key}.
const microgen-cote = require('microgen-cote');
const purchaseRequester = new microgen-cote.Requester({
name: 'Purchase Requester',
key: 'purchase',
});
const inventoryRequester = new microgen-cote.Requester({
name: 'Inventory Requester',
key: 'inventory',
});
Unlike environments, keys can't be used as an environment variable or part
of microgen-cote's configuration, but rather, should be provided as part of the first
argument to a component.
microgen-cote includes a Sockend component that provides a direct channel to
the frontend. This is extremely powerful and with power, comes great
responsibility. Exposing all the Responders and Publishers in the backend
to your frontend application probably isn't a good idea. Therefore microgen-cote offers
namespaces, which map conveniently to socket.io namespaces.
To help increase the security of backend services, components with
different namespaces won't recognize each other and try to communicate. This
effectively segregates the front-facing components. In order to allow a
component to talk to the frontend, you should use a namespace which shields
that service from the rest of the system. By incorporating multiple components
in a single service, you can basically create proxies and let your front-facing
components interact with the rest of the system in a secure way.
front-facing-service.js
const microgen-cote = require('microgen-cote');
const responder = new microgen-cote.Responder({
name: 'Conversion Sockend Responder',
namespace: 'conversion',
});
const conversionRequester = new microgen-cote.Requester({
name: 'Conversion Requester',
key: 'conversion backend',
});
responder.on('convert', (req, cb) => {
conversionRequester.send(req.type, req, cb); // proxy the request
});
backend-service.js
const microgen-cote = require('microgen-cote');
const responder = new microgen-cote.Responder({
name: 'Conversion Responder',
key: 'conversion backend',
});
const rates = { usd_eur: 0.91, eur_usd: 1.10 };
responder.on('convert', (req, cb) => {
cb(null, req.amount * rates[`${req.from}_${req.to}`]);
});
Just like keys, namespaces can also only be utilized as part of the first
argument to a component.
microgen-cote works either with IP multicast or IP broadcast, defaulting to broadcast. If
you wish to use multicast instead, you can pass in a multicast property with
the configuration object to microgen-cote. This will make sure that the discovery will
happen only with the given configuration.
In fact, this is the best way to segregate services, not in the application layer but at the network layer. This will create the minimal number of gossip messages and the biggest gains in terms of performance. Therefore, using different multicast addresses is better than using different environments or keys.
Much like environments, multicast addresses can be specified either as an
environment variable or as part of the main configuration object to the microgen-cote
require's. They can also be given as part of the second configuration object.
As an environment variable:
COTE_MULTICAST_ADDRESS=239.1.11.111 node service.js
As part of microgen-cote's module configuration:
const microgen-cote = require('microgen-cote')({ multicast: '239.1.11.111' });
As part of each component's discovery configuration:
const microgen-cote = require('microgen-cote');
const req = new microgen-cote.Requester({ name: 'req' }, { multicast: '239.1.11.111' });
While multicast is good for segmentation, certain scenarios may require the configuration be done over IP broadcast. In that case, broadcast address configuration helps. Much like multicast configuration, microgen-cote supports 3 different ways of supplying broadcast configuration.
Multicast configuration has precedence over broadcast. Therefore, when both configurations are applied, broadcast configuration will be ignored and multicast configuration will take over.
Also, microgen-cote uses broadcast by default. Hence, if no configuration is provided,
the broadcast address will be set to 255.255.255.255. If you want to use
broadcast, but have a different broadcast IP, you should configure it as shown
below.
As an environment variable:
COTE_BROADCAST_ADDRESS=255.255.255.255 node service.js
As part of microgen-cote's module configuration:
const microgen-cote = require('microgen-cote')({ broadcast: '255.255.255.255' });
As part of each component's discovery configuration:
const microgen-cote = require('microgen-cote');
const req = new microgen-cote.Requester({ name: 'req' }, { broadcast: '255.255.255.255' });
Here's a list of environment variables microgen-cote supports:
| Variable name | Description |
|---|---|
COTE_ENV | See Environments. |
COTE_MULTICAST_ADDRESS | See Multicast address. |
COTE_BROADCAST_ADDRESS | See Broadcast address. |
COTE_USE_HOST_NAMES | In certain, extremely rare conditions, auto-discovery might fail due to components reporting wrong IP addresses. If you find out that is the case, you can command microgen-cote to use the reported host names instead. |
COTE_DISCOVERY_REDIS | See Using centralized discovery tools. |
COTE_DISCOVERY_REDIS_URL | See Using centralized discovery tools. |
COTE_DISCOVERY_REDIS_HOST | See Using centralized discovery tools. |
DISCOVERY_HOSTNAME | See Using centralized discovery tools. |
COTE_REQUEST_TIMEOUT | See Requester Timeout. |
COTE_LOG | Boolean. Whether to display hello and status logs for other discovered services. Has precedence over COTE_STATUS_LOGS_ENABLED and COTE_HELLO_LOGS_ENABLED. |
COTE_HELLO_LOGS_ENABLED | Boolean. Whether to display hello logs from other discovered services. |
COTE_STATUS_LOGS_ENABLED | Boolean. Whether to display status logs from other discovered services. Has precedence over COTE_HELLO_LOGS_ENABLED. |
COTE_LOG_UNKNOWN_EVENTS | Boolean. Whether to log a message when a responder or subscriber receives an event that it has no listeners for. Defaults to true. |
COTE_CHECK_INTERVAL | Integer. The interval for checking if a discovered service has sent a heartbeat since the last check. |
COTE_HELLO_INTERVAL | Integer. The interval for sending a heartbeat hello signal. Should be less than COTE_CHECK_INTERVAL. |
COTE_NODE_TIMEOUT | Integer. The timeout duration that determines if a service is unreachable and thus removed. Should be greater than COTE_CHECK_INTERVAL. |
COTE_IGNORE_PROCESS | Boolean. Whether the services defined in this process should ignore other services from this process. This might be useful in a high-availability setup where one wants to enforce collaboration of services over the network, instead of local services within each process. |
microgen-cote plays extremely well with Docker Cloud. Even if your cloud provider doesn't support IP broadcast or multicast, you can still have the same functionality with Docker Cloud's Weave overlay networks.
Just deploy your microgen-cote applications just like any other Node.js application and even when your containers run in different machines on different continents, as long as they share an overlay network — which Docker Cloud assigns by default anyway — everything will work as expected.
Make sure to check out the e-commerce case study that implements a complete e-commerce application with microservices using microgen-cote. It features example Dockerfiles and docker-compose configurations in addition to Docker Cloud configurations.
It also has a Docker Swarm configuration to get you started on using microgen-cote with Docker Swarm, in any cloud environment.
microgen-cote is built to be zero-configuration, and relies on IP broadcast/multicast to work. cloud providers don't support this functionality (and they won't) out of the box. In these cases, one can use the Weave network overlay integration. However, this may not be suitable for everyone, due to varying reasons.
In these cases, in order to let microgen-cote work, we developed a plugin mechanism to accommodate different solutions that can serve as the automated service discovery tool. Currently, redis is supported out of the box, and microgen-cote makes use of the node_redis library, in case you want to use redis as the central discovery tool. If you need to use anything other than redis, please open a new issue and we may be able to help.
You should also set DISCOVERY_HOSTNAME to the IP address
of the container/instance since it defaults to machine's hostname which in
most cloud/docker setups is not routable.
microgen-cote aims to be as zero-conf as possible. Therefore, the discovery backend should be invisible to the developer. Since IP broadcast/multicast functionality is environment-specific, it makes sense to configure a centralized solution via environment variables as well. This way, the container deployment configurations such as Docker Swarm stack definitions can make use of the additional redis backend functionality, while developers can still use IP broadcast/multicast locally, with the same source code.
That's why microgen-cote uses environment variables that start with
COTE_DISCOVERY_REDIS. microgen-cote transforms any environment variable that
starts with COTE_DISCOVERY_REDIS to proper configuration for the
node_redis library. For example,
COTE_DISCOVERY_REDIS_URL=redis becomes { url: 'redis' } and
COTE_DISCOVERY_REDIS_HOST=redis COTE_DISCOVERY_REDIS_PORT=6379 becomes
{ host: 'redis', port: '6379' }.
| Variable name | Description |
|---|---|
COTE_DISCOVERY_REDIS | If you are running redis on localhost, setting this variable to true will use the locally available redis at port 6379. If you need any other redis URL or host, you don't need to use this variable. |
COTE_DISCOVERY_REDIS_URL | Sets the redis connection URL. Has to start with either redis:// or //. Enables the redis plugin. |
COTE_DISCOVERY_REDIS_HOST | Sets the redis connection host name. Enables the redis plugin. |
COTE_DISCOVERY_REDIS_PORT | Sets the redis connection port. Enables the redis plugin. |
DISCOVERY_HOSTNAME | This defaults to your machine's hostname. If this is not routable you need to set this to the routable IP address of this instance. |
microgen-cote also supports other connection options supported by node_redis in the same manner.
As an environment variable:
COTE_DISCOVERY_REDIS_HOST=redis DISCOVERY_HOSTNAME=127.0.0.1 node service.js
As part of microgen-cote's module configuration:
const microgen-cote = require('microgen-cote')({ redis: { host: 'redis' } });
As part of each component's discovery configuration:
const microgen-cote = require('microgen-cote');
const req = new microgen-cote.Requester({ name: 'req' }, { redis: { host: 'redis' } });
microgen-cote is battle-tested, solid and has been running in production across thousands of services since its inception in 2013. microgen-cote follows Semantic Versioning and although it's production-ready, we haven't released a version 1.0.0 yet. Although microgen-cote added many features in time, there hasn't been a single breaking API change since the beginning, so we didn't need to update the major version.
Make sure you don't run any of your services in cluster mode. It messes up the service discovery since it tries to load balance the UDP ports used internally for this purpose.
To use Cote properly within PM2 cluster mode, server instances should only be instantiated once.
To do so, utilize process.env.pm_id, which will return a value between 0 and the total number of instances(N). For example, if 10 instances of your app are running in cluster mode pm2 start app.js -i 10, process.env.pm_id will return a value of (0-9) inclusively.
// In thise case, we choose only the third app instance (2 because it is zero based) to instantiate a "SERVER"
// any number from 0 through 9 can be used, instead of 2
if (process.env.pm_id == 2) {
const microgen-cote = require('microgen-cote');
const timeService = new microgen-cote.Responder({
name: 'Time Service'
});
timeService.on('time', (req, cb) => {
cb(new Date());
});
}
Most cloud providers block IP broadcast and multicast, therefore you can't run microgen-cote in a multi-host environment without special software for an overlay network. For this purpose, Docker is the best tool. Deploy your application in Docker containers and you can take advantage of its overlay networks. Users of Docker Swarm can make use of the Weave Net plugin. Weave also has an addon for enabling multicast/broadcast for Kubernetes.
If you find the solutions with Docker Swarm and Kubernetes to be hard to get started with, you can use redis as a centralized discovery tool. Check out Using centralized discovery tools to see how you can set up redis to work with microgen-cote.
microgen-cote is under constant development, and has several important issues still open. We would therefore heavily appreciate if you headed to the project to see where we are in the development, picked an issue of your taste and gave us a hand.
If you would like to see a feature implemented or want to contribute a new feature, you are welcome to open an issue to discuss it and we will be more than happy to help.
If you choose to make a contribution, please fork this repository, work on a feature and submit a pull request. microgen-cote is the next level of microservices — be part of the revolution.
Copyright (c) 2013 Armagan Amcalar
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
FAQs
A Node.js library for building zero-configuration microservices.
The npm package microgen-cote receives a total of 8 weekly downloads. As such, microgen-cote popularity was classified as not popular.
We found that microgen-cote demonstrated a not healthy version release cadence and project activity because the last version was released a year ago. It has 1 open source maintainer collaborating on the project.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Security News
In this segment of the Risky Business podcast, Feross Aboukhadijeh and Patrick Gray discuss the challenges of tracking malware discovered in open source softare.
Research
Security News
A threat actor's playbook for exploiting the npm ecosystem was exposed on the dark web, detailing how to build a blockchain-powered botnet.
Security News
NVD’s backlog surpasses 20,000 CVEs as analysis slows and NIST announces new system updates to address ongoing delays.