Scaling Meteor (Q from Non-Technical Person)

Scaling Meteor (Q from Non-Technical Person) - javascript

TL;DR is it unwise to build a web app using Meteor as our full-stack framework when expecting lots of concurrent users at scale?
We are looking to build a web app that in its first phase, is a product hunt for music (feeding user submitted URL's from Youtube's API for play requests).
The benefits of Meteor seem to how quick it is to develop an MVP/prototype for learning about user behavior, but the risk seems to be in banking on using what is built during this phase throughout the course of the products life.
Wondering if anyone can help my ignorant brain understand in laymen's terms if anything I've asserted above is incorrect/correct, and if so, why? Sincerely grateful for any/all input on this

The communication and CPU overhead for a meteor app without accounts or subscriptions is nearly zero. However, as I point out in the answer to this question, the real resource constraints come from maintaining query result sets between the server and the clients. In other words, scaling becomes hard as subscriptions increase in number, size, and complexity.
Without knowing more about your product and how it works, my general recommendation is this: go for it, because meteor is going to give you a quick MVP. If you find you have a difficult time scaling because of a massive onslaught of users (congrats!), then you can always reduce your subscription expense by using a variety of tricks including non-reactive data (method calls).
Recommended reading: Scaling Meteor: The Challenges of Real-time Apps

Not unwise at all. There's a website called ClassCraft built using Meteor, that I think gets 15,000 logins a day.
Your easiest solution would probably be to go with a Node-based hosting service such as Modulus which will handle scaling automatically for you.
The next step towards a more DIY solution would be to just host your MongoDB on a provider such as Compose or MongoLab, but host the Meteor app itself on your own server(s). Look into mup for a simple deployment script, or Arunoda's newer solution he's working on, mupx, which uses Docker to deploy Meteor apps.
This just scratches the surface, but hopefully gives you a little more confidence that it has been done and it can be done.

Related

Confused about nodes purpose [duplicate]

Locked. This question and its answers are locked because the question is off-topic but has historical significance. It is not currently accepting new answers or interactions.
I am new to this kind of stuff, but lately I've been hearing a lot about how good Node.js is. Considering how much I love working with jQuery and JavaScript in general, I can't help but wonder how to decide when to use Node.js. The web application I have in mind is something like Bitly - takes some content, archives it.
From all the homework I have been doing in the last few days, I obtained the following information. Node.js
is a command-line tool that can be run as a regular web server and lets one run JavaScript programs
utilizes the great V8 JavaScript engine
is very good when you need to do several things at the same time
is event-based so all the wonderful Ajax-like stuff can be done on the server side
lets us share code between the browser and the backend
lets us talk with MySQL
Some of the sources that I have come across are:
Diving into Node.js – Introduction and Installation
Understanding NodeJS
Node by Example (Archive.is)
Let’s Make a Web App: NodePad
Considering that Node.js can be run almost out-of-the-box on Amazon's EC2 instances, I am trying to understand what type of problems require Node.js as opposed to any of the mighty kings out there like PHP, Python and Ruby. I understand that it really depends on the expertise one has on a language, but my question falls more into the general category of: When to use a particular framework and what type of problems is it particularly suited for?

You did a great job of summarizing what's awesome about Node.js. My feeling is that Node.js is especially suited for applications where you'd like to maintain a persistent connection from the browser back to the server. Using a technique known as "long-polling", you can write an application that sends updates to the user in real time. Doing long polling on many of the web's giants, like Ruby on Rails or Django, would create immense load on the server, because each active client eats up one server process. This situation amounts to a tarpit attack. When you use something like Node.js, the server has no need of maintaining separate threads for each open connection.
This means you can create a browser-based chat application in Node.js that takes almost no system resources to serve a great many clients. Any time you want to do this sort of long-polling, Node.js is a great option.
It's worth mentioning that Ruby and Python both have tools to do this sort of thing (eventmachine and twisted, respectively), but that Node.js does it exceptionally well, and from the ground up. JavaScript is exceptionally well situated to a callback-based concurrency model, and it excels here. Also, being able to serialize and deserialize with JSON native to both the client and the server is pretty nifty.
I look forward to reading other answers here, this is a fantastic question.
It's worth pointing out that Node.js is also great for situations in which you'll be reusing a lot of code across the client/server gap. The Meteor framework makes this really easy, and a lot of folks are suggesting this might be the future of web development. I can say from experience that it's a whole lot of fun to write code in Meteor, and a big part of this is spending less time thinking about how you're going to restructure your data, so the code that runs in the browser can easily manipulate it and pass it back.
Here's an article on Pyramid and long-polling, which turns out to be very easy to set up with a little help from gevent: TicTacToe and Long Polling with Pyramid.

I believe Node.js is best suited for real-time applications: online games, collaboration tools, chat rooms, or anything where what one user (or robot? or sensor?) does with the application needs to be seen by other users immediately, without a page refresh.
I should also mention that Socket.IO in combination with Node.js will reduce your real-time latency even further than what is possible with long polling. Socket.IO will fall back to long polling as a worst case scenario, and instead use web sockets or even Flash if they are available.
But I should also mention that just about any situation where the code might block due to threads can be better addressed with Node.js. Or any situation where you need the application to be event-driven.
Also, Ryan Dahl said in a talk that I once attended that the Node.js benchmarks closely rival Nginx for regular old HTTP requests. So if we build with Node.js, we can serve our normal resources quite effectively, and when we need the event-driven stuff, it's ready to handle it.
Plus it's all JavaScript all the time. Lingua Franca on the whole stack.

Reasons to use NodeJS:
It runs Javascript, so you can use the same language on server and client, and even share some code between them (e.g. for form validation, or to render views at either end.)
The single-threaded event-driven system is fast even when handling lots of requests at once, and also simple, compared to traditional multi-threaded Java or ROR frameworks.
The ever-growing pool of packages accessible through NPM, including client and server-side libraries/modules, as well as command-line tools for web development. Most of these are conveniently hosted on github, where sometimes you can report an issue and find it fixed within hours! It's nice to have everything under one roof, with standardized issue reporting and easy forking.
It has become the defacto standard environment in which to run Javascript-related tools and other web-related tools, including task runners, minifiers, beautifiers, linters, preprocessors, bundlers and analytics processors.
It seems quite suitable for prototyping, agile development and rapid product iteration.
Reasons not to use NodeJS:
It runs Javascript, which has no compile-time type checking. For large, complex safety-critical systems, or projects including collaboration between different organizations, a language which encourages contractual interfaces and provides static type checking may save you some debugging time (and explosions) in the long run. (Although the JVM is stuck with null, so please use Haskell for your nuclear reactors.)
Added to that, many of the packages in NPM are a little raw, and still under rapid development. Some libraries for older frameworks have undergone a decade of testing and bugfixing, and are very stable by now. Npmjs.org has no mechanism to rate packages, which has lead to a proliferation of packages doing more or less the same thing, out of which a large percentage are no longer maintained.
Nested callback hell. (Of course there are 20 different solutions to this...)
The ever-growing pool of packages can make one NodeJS project appear radically different from the next. There is a large diversity in implementations due to the huge number of options available (e.g. Express/Sails.js/Meteor/Derby). This can sometimes make it harder for a new developer to jump in on a Node project. Contrast that with a Rails developer joining an existing project: he should be able to get familiar with the app pretty quickly, because all Rails apps are encouraged to use a similar structure.
Dealing with files can be a bit of a pain. Things that are trivial in other languages, like reading a line from a text file, are weird enough to do with Node.js that there's a StackOverflow question on that with 80+ upvotes. There's no simple way to read one record at a time from a CSV file. Etc.
I love NodeJS, it is fast and wild and fun, but I am concerned it has little interest in provable-correctness. Let's hope we can eventually merge the best of both worlds. I am eager to see what will replace Node in the future... :)

To make it short:
Node.js is well suited for applications that have a lot of concurrent connections and each request only needs very few CPU cycles, because the event loop (with all the other clients) is blocked during execution of a function.
A good article about the event loop in Node.js is Mixu's tech blog: Understanding the node.js event loop.

I have one real-world example where I have used Node.js. The company where I work got one client who wanted to have a simple static HTML website. This website is for selling one item using PayPal and the client also wanted to have a counter which shows the amount of sold items. Client expected to have huge amount of visitors to this website. I decided to make the counter using Node.js and the Express.js framework.
The Node.js application was simple. Get the sold items amount from a Redis database, increase the counter when item is sold and serve the counter value to users via the API.
Some reasons why I chose to use Node.js in this case
It is very lightweight and fast. There has been over 200000 visits on this website in three weeks and minimal server resources has been able to handle it all.
The counter is really easy to make to be real time.
Node.js was easy to configure.
There are lots of modules available for free. For example, I found a Node.js module for PayPal.
In this case, Node.js was an awesome choice.

The most important reasons to start your next project using Node ...
All the coolest dudes are into it ... so it must be fun.
You can hangout at the cooler and have lots of Node adventures to brag about.
You're a penny pincher when it comes to cloud hosting costs.
Been there done that with Rails
You hate IIS deployments
Your old IT job is getting rather dull and you wish you were in a shiny new Start Up.
What to expect ...
You'll feel safe and secure with Express without all the server bloatware you never needed.
Runs like a rocket and scales well.
You dream it. You installed it. The node package repo npmjs.org is the largest ecosystem of open source libraries in the world.
Your brain will get time warped in the land of nested callbacks ...
... until you learn to keep your Promises.
Sequelize and Passport are your new API friends.
Debugging mostly async code will get umm ... interesting .
Time for all Noders to master Typescript.
Who uses it?
PayPal, Netflix, Walmart, LinkedIn, Groupon, Uber, GoDaddy, Dow Jones
Here's why they switched to Node.

There is nothing like Silver Bullet. Everything comes with some cost associated with it. It is like if you eat oily food, you will compromise your health and healthy food does not come with spices like oily food. It is individual choice whether they want health or spices as in their food.
Same way Node.js consider to be used in specific scenario. If your app does not fit into that scenario you should not consider it for your app development. I am just putting my thought on the same:
When to use Node.JS
If your server side code requires very few cpu cycles. In other world you are doing non blocking operation and does not have heavy algorithm/Job which consumes lots of CPU cycles.
If you are from Javascript back ground and comfortable in writing Single Threaded code just like client side JS.
When NOT to use Node.JS
Your server request is dependent on heavy CPU consuming algorithm/Job.
Scalability Consideration with Node.JS
Node.JS itself does not utilize all core of underlying system and it is single threaded by default, you have to write logic by your own to utilize multi core processor and make it multi threaded.
Node.JS Alternatives
There are other option to use in place of Node.JS however Vert.x seems to be pretty promising and has lots of additional features like polygot and better scalability considerations.

Another great thing that I think no one has mentioned about Node.js is the amazing community, the package management system (npm) and the amount of modules that exist that you can include by simply including them in your package.json file.

My piece: nodejs is great for making real time systems like analytics, chat-apps, apis, ad servers, etc.
Hell, I made my first chat app using nodejs and socket.io under 2 hours and that too during exam
week!
Edit
Its been several years since I have started using nodejs and I have used it in making many different things including static file servers, simple analytics, chat apps and much more.
This is my take on when to use nodejs
When to use
When making system which put emphasis on concurrency and speed.
Sockets only servers like chat apps, irc apps, etc.
Social networks which put emphasis on realtime resources like geolocation, video stream, audio stream, etc.
Handling small chunks of data really fast like an analytics webapp.
As exposing a REST only api.
When not to use
Its a very versatile webserver so you can use it wherever you want but probably not these places.
Simple blogs and static sites.
Just as a static file server.
Keep in mind that I am just nitpicking. For static file servers, apache is better mainly because it is widely available. The nodejs community has grown larger and more mature over the years and it is safe to say nodejs can be used just about everywhere if you have your own choice of hosting.

It can be used where
Applications that are highly event driven & are heavily I/O bound
Applications handling a large number of connections to other systems
Real-time applications (Node.js was designed from the ground up for real time and to be easy
to use.)
Applications that juggle scads of information streaming to and from other sources
High traffic, Scalable applications
Mobile apps that have to talk to platform API & database, without having to do a lot of data
analytics
Build out networked applications
Applications that need to talk to the back end very often
On Mobile front, prime-time companies have relied on Node.js for their mobile solutions. Check out why?
LinkedIn is a prominent user. Their entire mobile stack is built on Node.js. They went from running 15 servers with 15 instances on each physical machine, to just 4 instances – that can handle double the traffic!
eBay launched ql.io, a web query language for HTTP APIs, which uses Node.js as the runtime stack. They were able to tune a regular developer-quality Ubuntu workstation to handle more than 120,000 active connections per node.js process, with each connection consuming about 2kB memory!
Walmart re-engineered its mobile app to use Node.js and pushed its JavaScript processing to the server.
Read more at: http://www.pixelatingbits.com/a-closer-look-at-mobile-app-development-with-node-js/

Node best for concurrent request handling -
So, Let’s start with a story. From last 2 years I am working on JavaScript and developing web front end and I am enjoying it. Back end guys provide’s us some API’s written in Java,python (we don’t care) and we simply write a AJAX call, get our data and guess what ! we are done. But in real it is not that easy, If data we are getting is not correct or there is some server error then we stuck and we have to contact our back end guys over the mail or chat(sometimes on whatsApp too :).) This is not cool. What if we wrote our API’s in JavaScript and call those API’s from our front end ? Yes that’s pretty cool because if we face any problem in API we can look into it. Guess what ! you can do this now , How ? – Node is there for you.
Ok agreed that you can write your API in JavaScript but what if I am ok with above problem. Do you have any other reason to use node for rest API ?
so here is the magic begins. Yes I do have other reasons to use node for our API’s.
Let’s go back to our traditional rest API system which is based on either blocking operation or threading. Suppose two concurrent request occurs( r1 and r2) , each of them require database operation. So In traditional system what will happens :
1. Waiting Way : Our server starts serving r1 request and waits for query response. after completion of r1 , server starts to serve r2 and does it in same way. So waiting is not a good idea because we don’t have that much time.
2. Threading Way : Our server will creates two threads for both requests r1 and r2 and serve their purpose after querying database so cool its fast.But it is memory consuming because you can see we started two threads also problem increases when both request is querying same data then you have to deal with deadlock kind of issues . So its better than waiting way but still issues are there.
Now here is , how node will do it:
3. Nodeway : When same concurrent request comes in node then it will register an event with its callback and move ahead it will not wait for query response for a particular request.So when r1 request comes then node’s event loop (yes there is an event loop in node which serves this purpose.) register an event with its callback function and move ahead for serving r2 request and similarly register its event with its callback. Whenever any query finishes it triggers its corresponding event and execute its callback to completion without being interrupted.
So no waiting, no threading , no memory consumption – yes this is nodeway for serving rest API.

My one more reason to choose Node.js for a new project is:
Be able to do pure cloud based development
I have used Cloud9 IDE for a while and now I can't imagine without it, it covers all the development lifecycles. All you need is a browser and you can code anytime anywhere on any devices. You don't need to check in code in one Computer(like at home), then checkout in another computer(like at work place).
Of course, there maybe cloud based IDE for other languages or platforms (Cloud 9 IDE is adding supports for other languages as well), but using Cloud 9 to do Node.js developement is really a great experience for me.

One more thing node provides is the ability to create multiple v8 instanes of node using node's child process( childProcess.fork() each requiring 10mb memory as per docs) on the fly, thus not affecting the main process running the server. So offloading a background job that requires huge server load becomes a child's play and we can easily kill them as and when needed.
I've been using node a lot and in most of the apps we build, require server connections at the same time thus a heavy network traffic. Frameworks like Express.js and the new Koajs (which removed callback hell) have made working on node even more easier.

Donning asbestos longjohns...
Yesterday my title with Packt Publications, Reactive Programming with JavaScript. It isn't really a Node.js-centric title; early chapters are intended to cover theory, and later code-heavy chapters cover practice. Because I didn't really think it would be appropriate to fail to give readers a webserver, Node.js seemed by far the obvious choice. The case was closed before it was even opened.
I could have given a very rosy view of my experience with Node.js. Instead I was honest about good points and bad points I encountered.
Let me include a few quotes that are relevant here:
Warning: Node.js and its ecosystem are hot--hot enough to burn you badly!
When I was a teacher’s assistant in math, one of the non-obvious suggestions I was told was not to tell a student that something was “easy.” The reason was somewhat obvious in retrospect: if you tell people something is easy, someone who doesn’t see a solution may end up feeling (even more) stupid, because not only do they not get how to solve the problem, but the problem they are too stupid to understand is an easy one!
There are gotchas that don’t just annoy people coming from Python / Django, which immediately reloads the source if you change anything. With Node.js, the default behavior is that if you make one change, the old version continues to be active until the end of time or until you manually stop and restart the server. This inappropriate behavior doesn’t just annoy Pythonistas; it also irritates native Node.js users who provide various workarounds. The StackOverflow question “Auto-reload of files in Node.js” has, at the time of this writing, over 200 upvotes and 19 answers; an edit directs the user to a nanny script, node-supervisor, with homepage at http://tinyurl.com/reactjs-node-supervisor. This problem affords new users with great opportunity to feel stupid because they thought they had fixed the problem, but the old, buggy behavior is completely unchanged. And it is easy to forget to bounce the server; I have done so multiple times. And the message I would like to give is, “No, you’re not stupid because this behavior of Node.js bit your back; it’s just that the designers of Node.js saw no reason to provide appropriate behavior here. Do try to cope with it, perhaps taking a little help from node-supervisor or another solution, but please don’t walk away feeling that you’re stupid. You’re not the one with the problem; the problem is in Node.js’s default behavior.”
This section, after some debate, was left in, precisely because I don't want to give an impression of “It’s easy.” I cut my hands repeatedly while getting things to work, and I don’t want to smooth over difficulties and set you up to believe that getting Node.js and its ecosystem to function well is a straightforward matter and if it’s not straightforward for you too, you don’t know what you’re doing. If you don’t run into obnoxious difficulties using Node.js, that’s wonderful. If you do, I would hope that you don’t walk away feeling, “I’m stupid—there must be something wrong with me.” You’re not stupid if you experience nasty surprises dealing with Node.js. It’s not you! It’s Node.js and its ecosystem!
The Appendix, which I did not really want after the rising crescendo in the last chapters and the conclusion, talks about what I was able to find in the ecosystem, and provided a workaround for moronic literalism:
Another database that seemed like a perfect fit, and may yet be redeemable, is a server-side implementation of the HTML5 key-value store. This approach has the cardinal advantage of an API that most good front-end developers understand well enough. For that matter, it’s also an API that most not-so-good front-end developers understand well enough. But with the node-localstorage package, while dictionary-syntax access is not offered (you want to use localStorage.setItem(key, value) or localStorage.getItem(key), not localStorage[key]), the full localStorage semantics are implemented, including a default 5MB quota—WHY? Do server-side JavaScript developers need to be protected from themselves?
For client-side database capabilities, a 5MB quota per website is really a generous and useful amount of breathing room to let developers work with it. You could set a much lower quota and still offer developers an immeasurable improvement over limping along with cookie management. A 5MB limit doesn’t lend itself very quickly to Big Data client-side processing, but there is a really quite generous allowance that resourceful developers can use to do a lot. But on the other hand, 5MB is not a particularly large portion of most disks purchased any time recently, meaning that if you and a website disagree about what is reasonable use of disk space, or some site is simply hoggish, it does not really cost you much and you are in no danger of a swamped hard drive unless your hard drive was already too full. Maybe we would be better off if the balance were a little less or a little more, but overall it’s a decent solution to address the intrinsic tension for a client-side context.
However, it might gently be pointed out that when you are the one writing code for your server, you don’t need any additional protection from making your database more than a tolerable 5MB in size. Most developers will neither need nor want tools acting as a nanny and protecting them from storing more than 5MB of server-side data. And the 5MB quota that is a golden balancing act on the client-side is rather a bit silly on a Node.js server. (And, for a database for multiple users such as is covered in this Appendix, it might be pointed out, slightly painfully, that that’s not 5MB per user account unless you create a separate database on disk for each user account; that’s 5MB shared between all user accounts together. That could get painful if you go viral!) The documentation states that the quota is customizable, but an email a week ago to the developer asking how to change the quota is unanswered, as was the StackOverflow question asking the same. The only answer I have been able to find is in the Github CoffeeScript source, where it is listed as an optional second integer argument to a constructor. So that’s easy enough, and you could specify a quota equal to a disk or partition size. But besides porting a feature that does not make sense, the tool’s author has failed completely to follow a very standard convention of interpreting 0 as meaning “unlimited” for a variable or function where an integer is to specify a maximum limit for some resource use. The best thing to do with this misfeature is probably to specify that the quota is Infinity:
if (typeof localStorage === 'undefined' || localStorage === null)
{
var LocalStorage = require('node-localstorage').LocalStorage;
localStorage = new LocalStorage(__dirname + '/localStorage',
Infinity);
}
Swapping two comments in order:
People needlessly shot themselves in the foot constantly using JavaScript as a whole, and part of JavaScript being made respectable language was a Douglas Crockford saying in essence, “JavaScript as a language has some really good parts and some really bad parts. Here are the good parts. Just forget that anything else is there.” Perhaps the hot Node.js ecosystem will grow its own “Douglas Crockford,” who will say, “The Node.js ecosystem is a coding Wild West, but there are some real gems to be found. Here’s a roadmap. Here are the areas to avoid at almost any cost. Here are the areas with some of the richest paydirt to be found in ANY language or environment.”
Perhaps someone else can take those words as a challenge, and follow Crockford’s lead and write up “the good parts” and / or “the better parts” for Node.js and its ecosystem. I’d buy a copy!
And given the degree of enthusiasm and sheer work-hours on all projects, it may be warranted in a year, or two, or three, to sharply temper any remarks about an immature ecosystem made at the time of this writing. It really may make sense in five years to say, “The 2015 Node.js ecosystem had several minefields. The 2020 Node.js ecosystem has multiple paradises.”

If your application mainly tethers web apis, or other io channels, give or take a user interface, node.js may be a fair pick for you, especially if you want to squeeze out the most scalability, or, if your main language in life is javascript (or javascript transpilers of sorts). If you build microservices, node.js is also okay. Node.js is also suitable for any project that is small or simple.
Its main selling point is it allows front-enders take responsibility for back-end stuff rather than the typical divide. Another justifiable selling point is if your workforce is javascript oriented to begin with.
Beyond a certain point however, you cannot scale your code without terrible hacks for forcing modularity, readability and flow control. Some people like those hacks though, especially coming from an event-driven javascript background, they seem familiar or forgivable.
In particular, when your application needs to perform synchronous flows, you start bleeding over half-baked solutions that slow you down considerably in terms of your development process. If you have computation intensive parts in your application, tread with caution picking (only) node.js. Maybe http://koajs.com/ or other novelties alleviate those originally thorny aspects, compared to when I originally used node.js or wrote this.

I can share few points where&why to use node js.
For realtime applications like chat,collaborative editing better we go with nodejs as it is event base where fire event and data to clients from server.
Simple and easy to understand as it is javascript base where most of people have idea.
Most of current web applications going towards angular js&backbone, with node it is easy to interact with client side code as both will use json data.
Lot of plugins available.
Drawbacks:-
Node will support most of databases but best is mongodb which won't support complex joins and others.
Compilation Errors...developer should handle each and every exceptions other wise if any error accord application will stop working where again we need to go and start it manually or using any automation tool.
Conclusion:-
Nodejs best to use for simple and real time applications..if you have very big business logic and complex functionality better should not use nodejs.
If you want to build an application along with chat and any collaborative functionality.. node can be used in specific parts and remain should go with your convenience technology.

Node is great for quick prototypes but I'd never use it again for anything complex.
I spent 20 years developing a relationship with a compiler and I sure miss it.
Node is especially painful for maintaining code that you haven't visited for awhile. Type info and compile time error detection are GOOD THINGS. Why throw all that out? For what? And dang, when something does go south the stack traces quite often completely useless.

What is the best way I can scale my nodejs app?

The basics
Right now a few of my friends and I are trying to develope a browser game made in nodejs. It's a multiplayer top-down shooter, and most of both the client-side and server-side code is in javascript. We have a good general direction that we'd like to go in, and we're having a lot of fun developing the game. One of our goals when making this game was to make it as hard as possible to cheat. Do do that, we have all of the game logic handled server-side. The client only sends their input the the server via web socket, and the server updates the client (also web socket) with what is happening in the game. Here's the start of our problem.
All of the server side math is getting pretty hefty, and we're finding that we need to scale in some way to handle anything more than 10 players (we want to be able to host many more). At first we had figured that we could just scale vertically as we needed to, but since nodejs is single threaded, is can only take advantage of one core. This means that getting a beefier server won't help that problem. Our only solution is to scale horizontally.
Why we're asking here
We haven't been able to find any good examples of how to scale out a nodejs game. Our use case is pretty particular, and while we've done our best to do this by ourselves, we could really benefit from outside opinions and advice
Details
We've already put a LOT of thought into how to solve this problem. We've been working on it for over a week. Here's what we have put together so far:
Four types of servers
We're splitting tasks into 4 different 'types' of servers. Each one will have a specific task it completes.
The proxy server
The proxy server would sit at the front of the entire stack, and be the only server directly accessible from the internet (there could potentially be more of these). It would have haproxy on it, and it would route all connections to the web servers. We chose haproxy because of its rich feature set, reliability, and nearly unbeatable speed.
The web server
The web server would receive the web-requests, and serve all web-pages. They would also handle lobby creation/management and game creation/management. To do this, they would tell the game servers what lobbies it has, what users are in that lobby, and info about the game they're going to play. The web servers would then update the game servers about user input, and the game server would update the web servers (who would then update the clients) of what's happening in the game. The web servers would use TCP sockets to communicate with the game servers about any type of management, and they would use UDP sockets when communicating about game updates. This would all be done with nodejs.
The game server
The game server would handle all the game math and variable updates about the game. The game servers also communicate with the db servers to record cool stats about players in game. This would be done with nodejs.
The db server
The db server would host the database. This part actually turned out to be the easiest since we found rethinkdb, the coolest db ever. This scales easily, and oddly enough, turned out to be the easiest part of scaling our application.
Some other details
If you're having trouble getting your head around our whole getup, look at this, it's a semi-accurate chart of how we think we'll scale.
If you're just curious, or think it might be helpful to look at our game, it's currently hosted in it's un-scaled state here.
Some things we don't want
We don't want to use the cluster module of nodejs. It isn't stable (said here), and it doesn't scale to other servers, only other processors. We'd like to just take the leap to horizontal scaling.
Our question, summed up
We hope we're going in the right direction, and we've done our homework, but we're not certain. We could certainly take a few tips on how to do this the right way.
Thanks
I realize that this is a pretty long question, and making a well thought out answer will not be easy, but I would really appreciate it.
Thanks!!

Following my spontaneous thoughts on your case:
Multicore usage
node.js can scale with multiple cores as well. How, you can read for example here (or just think about it: You have one thread/process running on one core, what do you need to use multiple cores? Multiple threads or multiple processes. Push work from main thread to other threads or processes and you are done).
I personally would say it is childish to develop an application, which does not make use of multiple cores. If you make use of some background processes, ok, but if you until now only do work in the node.js main event loop, you should definitely invest some time to make the app scalable over cores.
Implementing something like IPC is not that easy by the way. You can do, but if your case is complicated maybe you are good to go with the cluster module. This is obviously not your favorite, but just because something is called "experimental" it does not mean it's trashy. Just give it a try, maybe you can even fix some bugs of the module on the way. It's most likely better to use some broadly used software for complex problems, than invent a new wheel.
You should also (if you do not already) think about (wise) usage of nextTick functionality. This allows the main event loop to pause some cpu intensive task and perform other work in the meanwhile. You can read about it for example here.
General thoughts on computations
You should definitely take a very close look at your algorithms of the game engine. You already noticed that this is your bottleneck right now and actually computations are the most critical part of mostly every game. Scaling does solve this problem in one way, but scaling introduces other problems. Also you cannot throw "scaling" as problem solver on everything and expect every problem to disappear.
Your best bet is to make your game code elegant and fast. Think about how to solve problems efficiently. If you cannot solve something in Javascript efficiently, but the problem can easily be extracted, why not write a little C component instead? This counts as a separate process as well, which reduces load on your main node.js event loop.
Proxy?
Personally I do not see the advantage of the proxy level right now. You do not seem to expect large amount of users, you therefore won't need to solve problems like CDN solves or whatever... it's okay to think about it, but I would not invest much time there right now.
Technically there is a high chance your webserver software provides proxy functionality anyway. So it is ok to have it on the paper, but I would not plan with dedicated hardware right now.
Epilogue
The rest seems more or less fine to me.

Little late to the game, but take a look here: http://goldfirestudios.com/blog/136/Horizontally-Scaling-Node.js-and-WebSockets-with-Redis
You did not mention anything to do with memory management. As you know, nodejs doesn't share its memory with other processes, so an in-memory database is a must if you want to scale. (Redis, Memcache, etc). You need to setup a publisher & subscriber event on each node to accept incoming requests from redis. This way, you can scale up x nilo amount of servers (infront of your HAProxy) and utilize the data piped from redis.
There is also this node addon: http://blog.varunajayasiri.com/shared-memory-with-nodejs That lets you share memory between processes, but only works under Linux. This will help if you don't want to send data across local processes all the time or have to deal with nodes ipc api.
You can also fork child processes within node for a new v8 isolate to help with expensive cpu bound tasks. For example, players can kill monsters and obtain quite a bit of loot within my action rpg game. I have a child process called LootGenerater, and basically whenever a player kills a monster it sends the game id, mob_id, and user_id to the process via the default IPC api .send. Once the child process receives it, it iterates over the large loot table and manages the items (stores to redis, or whatever) and pipes it back.
This helps free up the event loop greatly, and just one idea I can think of to help you scale. But most importantly you will want to use an in-memory database system and make sure your game code architecture is designed around whatever database system you use. Don't make the mistake I did by now having to re-write everything :)
Hope this helps!
Note: If you do decide to go with Memcache, you will need to utilize another pub/sub system.

how to refactor a manually deployed single instance app into a client deployed multi-instance or multi-tenant app

I have an meteor.js application that can be manually configured and deployed for a single instance quite nicely.
It is now time to refactor the application's architecture and build out the infrastructure around the app to allow it to be client deployed and update-able.
I'd like to have client come to a page where they can sign-up for the app, an instance or a tenancy would be automatically setup for them, and they could start using it. On the back-end there would be infrastructure to manage updates to the application.
There are some obvious decision that need to be made:
Do I refactor it to be multi-tenant? (more application code modifications)
Do I refactor it to be multi-instance? (more infrastructure build-out and code)
Is it a hybrid? (one application but multiple databases)
What tests does one apply to determine the correct answer to the above question? and what are the pros and cons of each?
Once that decision has been made, does there exist design patterns to guide or inspire a proper refactor, and/or what learning resources exist out there for someone who has not built a multi-tenant or multi-instance app?
If its multi-instancing should the instancing and updating be part of the application itself or is there another layer of code and tools that should be built to manage that part?

I count 3 questions here, and you might find it valuable to split them into separate threads. At any rate:
1) What test to determine the correct architecture? Well, a hard-assed look at how much it's going to cost to support each architecture vs. how quickly each can be implemented & how many waiting customers you have seems in order. Hard-assed because, frankly, you probably already have a preference, and unless you're willing to set that aside the answer I give here is moot. If this is for a business, remember that revenue rules -- without revenue even the most beautiful & elegant architecture is unimportant. With revenue, you can fix most architecture mistakes in time.
2) What are good design patterns for multi-tenant, embed-able applications? I'm not sure that design patterns are the right answer, but rather data management & testing rigor. The goal here is to ensure that Client A's customers will never get a hint of Client B's customer data, even if a single individual is a user of both Client A and Client B. Careful attention to API keys and session key management is the order of the day.
3) Instance management in app, or a separate tool? I'm going to go out on a limb, and suggest that nobody will be able to satisfactorily answer this question without an analysis of your current application and infrastructure. Maybe you have an application that is mostly self deploying, and only needs a few more lines to set up a new DB, or launch a new AWS instance, or whatever... Or maybe you have a highly manual process. This may also be influenced by your choice of architecture from Question 1, and/or how much time you have. See note about revenue from Question 1.

Is Meteor.js stable enough for production?

I was really amazed about's meteor.js features and ease of use, im really falling in love with it. I just wanted to know if you expert guys think it is stable enough for using it in production for a medium sized project.
Thanks for your advice!
Update Nov 2014:
Meteor 1.0 was just released, this is the first ready for production version finally, more on meteor check out this link

6 months with Meteor has the best answer to this question I've found so far.
Here are a few excerpts:
Meteor can’t be used for all real-world apps just yet. Meteor doesn’t have server-side rendering yet, so it’s not ideal for sites that need to load very fast (like e-commerce
sites) or work on underpowered devices (like older mobile phones).
...
So I would say right now Meteor will be a perfect choice for a few apps (anything that strongly depends on real-time interactions), a great choice for most of them, and a very bad one for a few specific cases.
It sounds like Meteor is definitely promising and particularly well-suited for real-time essential web apps, however it's not production ready for all web apps at the moment. At the time of writing this meteor is at version 0.7.0.1.

It's hard to answer this, because use cases vary so much.
I have an in-house application running on Meteor, which is basically a web-based SMS client with address book, fully integrated with our custom, in-house CRM. The Meteor app has 25 simultaneous users, and has been used to send or receive 70,000 SMS messages over the past 5 months, with total downtime measurable in minutes (which was due to network outages, not a problem with the application or framework).
There are definitely missing features, but assuming the features you need are implemented, the framework can be stable enough for production.

Really depends on what you're doing and how many users will connect to it. Using the current version 0.6.x you will need to tweak things that maybe Meteor devs are going to tweak, like MongoDB operations. Also, many packages are going to be released in future that will help a lot some work you are going to build yourself right now (for e.g. the IronRouter that C. Mather is building is very complex to build but very useful if you need complex operations in your app, like subscribing automatically in certain views, building forms etc)

How to decide when to use Node.js?

Locked. This question and its answers are locked because the question is off-topic but has historical significance. It is not currently accepting new answers or interactions.
I am new to this kind of stuff, but lately I've been hearing a lot about how good Node.js is. Considering how much I love working with jQuery and JavaScript in general, I can't help but wonder how to decide when to use Node.js. The web application I have in mind is something like Bitly - takes some content, archives it.
From all the homework I have been doing in the last few days, I obtained the following information. Node.js
is a command-line tool that can be run as a regular web server and lets one run JavaScript programs
utilizes the great V8 JavaScript engine
is very good when you need to do several things at the same time
is event-based so all the wonderful Ajax-like stuff can be done on the server side
lets us share code between the browser and the backend
lets us talk with MySQL
Some of the sources that I have come across are:
Diving into Node.js – Introduction and Installation
Understanding NodeJS
Node by Example (Archive.is)
Let’s Make a Web App: NodePad
Considering that Node.js can be run almost out-of-the-box on Amazon's EC2 instances, I am trying to understand what type of problems require Node.js as opposed to any of the mighty kings out there like PHP, Python and Ruby. I understand that it really depends on the expertise one has on a language, but my question falls more into the general category of: When to use a particular framework and what type of problems is it particularly suited for?

I believe Node.js is best suited for real-time applications: online games, collaboration tools, chat rooms, or anything where what one user (or robot? or sensor?) does with the application needs to be seen by other users immediately, without a page refresh.
I should also mention that Socket.IO in combination with Node.js will reduce your real-time latency even further than what is possible with long polling. Socket.IO will fall back to long polling as a worst case scenario, and instead use web sockets or even Flash if they are available.
But I should also mention that just about any situation where the code might block due to threads can be better addressed with Node.js. Or any situation where you need the application to be event-driven.
Also, Ryan Dahl said in a talk that I once attended that the Node.js benchmarks closely rival Nginx for regular old HTTP requests. So if we build with Node.js, we can serve our normal resources quite effectively, and when we need the event-driven stuff, it's ready to handle it.
Plus it's all JavaScript all the time. Lingua Franca on the whole stack.

To make it short:
Node.js is well suited for applications that have a lot of concurrent connections and each request only needs very few CPU cycles, because the event loop (with all the other clients) is blocked during execution of a function.
A good article about the event loop in Node.js is Mixu's tech blog: Understanding the node.js event loop.

I have one real-world example where I have used Node.js. The company where I work got one client who wanted to have a simple static HTML website. This website is for selling one item using PayPal and the client also wanted to have a counter which shows the amount of sold items. Client expected to have huge amount of visitors to this website. I decided to make the counter using Node.js and the Express.js framework.
The Node.js application was simple. Get the sold items amount from a Redis database, increase the counter when item is sold and serve the counter value to users via the API.
Some reasons why I chose to use Node.js in this case
It is very lightweight and fast. There has been over 200000 visits on this website in three weeks and minimal server resources has been able to handle it all.
The counter is really easy to make to be real time.
Node.js was easy to configure.
There are lots of modules available for free. For example, I found a Node.js module for PayPal.
In this case, Node.js was an awesome choice.

The most important reasons to start your next project using Node ...
All the coolest dudes are into it ... so it must be fun.
You can hangout at the cooler and have lots of Node adventures to brag about.
You're a penny pincher when it comes to cloud hosting costs.
Been there done that with Rails
You hate IIS deployments
Your old IT job is getting rather dull and you wish you were in a shiny new Start Up.
What to expect ...
You'll feel safe and secure with Express without all the server bloatware you never needed.
Runs like a rocket and scales well.
You dream it. You installed it. The node package repo npmjs.org is the largest ecosystem of open source libraries in the world.
Your brain will get time warped in the land of nested callbacks ...
... until you learn to keep your Promises.
Sequelize and Passport are your new API friends.
Debugging mostly async code will get umm ... interesting .
Time for all Noders to master Typescript.
Who uses it?
PayPal, Netflix, Walmart, LinkedIn, Groupon, Uber, GoDaddy, Dow Jones
Here's why they switched to Node.

There is nothing like Silver Bullet. Everything comes with some cost associated with it. It is like if you eat oily food, you will compromise your health and healthy food does not come with spices like oily food. It is individual choice whether they want health or spices as in their food.
Same way Node.js consider to be used in specific scenario. If your app does not fit into that scenario you should not consider it for your app development. I am just putting my thought on the same:
When to use Node.JS
If your server side code requires very few cpu cycles. In other world you are doing non blocking operation and does not have heavy algorithm/Job which consumes lots of CPU cycles.
If you are from Javascript back ground and comfortable in writing Single Threaded code just like client side JS.
When NOT to use Node.JS
Your server request is dependent on heavy CPU consuming algorithm/Job.
Scalability Consideration with Node.JS
Node.JS itself does not utilize all core of underlying system and it is single threaded by default, you have to write logic by your own to utilize multi core processor and make it multi threaded.
Node.JS Alternatives
There are other option to use in place of Node.JS however Vert.x seems to be pretty promising and has lots of additional features like polygot and better scalability considerations.

Another great thing that I think no one has mentioned about Node.js is the amazing community, the package management system (npm) and the amount of modules that exist that you can include by simply including them in your package.json file.

My piece: nodejs is great for making real time systems like analytics, chat-apps, apis, ad servers, etc.
Hell, I made my first chat app using nodejs and socket.io under 2 hours and that too during exam
week!
Edit
Its been several years since I have started using nodejs and I have used it in making many different things including static file servers, simple analytics, chat apps and much more.
This is my take on when to use nodejs
When to use
When making system which put emphasis on concurrency and speed.
Sockets only servers like chat apps, irc apps, etc.
Social networks which put emphasis on realtime resources like geolocation, video stream, audio stream, etc.
Handling small chunks of data really fast like an analytics webapp.
As exposing a REST only api.
When not to use
Its a very versatile webserver so you can use it wherever you want but probably not these places.
Simple blogs and static sites.
Just as a static file server.
Keep in mind that I am just nitpicking. For static file servers, apache is better mainly because it is widely available. The nodejs community has grown larger and more mature over the years and it is safe to say nodejs can be used just about everywhere if you have your own choice of hosting.

It can be used where
Applications that are highly event driven & are heavily I/O bound
Applications handling a large number of connections to other systems
Real-time applications (Node.js was designed from the ground up for real time and to be easy
to use.)
Applications that juggle scads of information streaming to and from other sources
High traffic, Scalable applications
Mobile apps that have to talk to platform API & database, without having to do a lot of data
analytics
Build out networked applications
Applications that need to talk to the back end very often
On Mobile front, prime-time companies have relied on Node.js for their mobile solutions. Check out why?
LinkedIn is a prominent user. Their entire mobile stack is built on Node.js. They went from running 15 servers with 15 instances on each physical machine, to just 4 instances – that can handle double the traffic!
eBay launched ql.io, a web query language for HTTP APIs, which uses Node.js as the runtime stack. They were able to tune a regular developer-quality Ubuntu workstation to handle more than 120,000 active connections per node.js process, with each connection consuming about 2kB memory!
Walmart re-engineered its mobile app to use Node.js and pushed its JavaScript processing to the server.
Read more at: http://www.pixelatingbits.com/a-closer-look-at-mobile-app-development-with-node-js/

Node best for concurrent request handling -
So, Let’s start with a story. From last 2 years I am working on JavaScript and developing web front end and I am enjoying it. Back end guys provide’s us some API’s written in Java,python (we don’t care) and we simply write a AJAX call, get our data and guess what ! we are done. But in real it is not that easy, If data we are getting is not correct or there is some server error then we stuck and we have to contact our back end guys over the mail or chat(sometimes on whatsApp too :).) This is not cool. What if we wrote our API’s in JavaScript and call those API’s from our front end ? Yes that’s pretty cool because if we face any problem in API we can look into it. Guess what ! you can do this now , How ? – Node is there for you.
Ok agreed that you can write your API in JavaScript but what if I am ok with above problem. Do you have any other reason to use node for rest API ?
so here is the magic begins. Yes I do have other reasons to use node for our API’s.
Let’s go back to our traditional rest API system which is based on either blocking operation or threading. Suppose two concurrent request occurs( r1 and r2) , each of them require database operation. So In traditional system what will happens :
1. Waiting Way : Our server starts serving r1 request and waits for query response. after completion of r1 , server starts to serve r2 and does it in same way. So waiting is not a good idea because we don’t have that much time.
2. Threading Way : Our server will creates two threads for both requests r1 and r2 and serve their purpose after querying database so cool its fast.But it is memory consuming because you can see we started two threads also problem increases when both request is querying same data then you have to deal with deadlock kind of issues . So its better than waiting way but still issues are there.
Now here is , how node will do it:
3. Nodeway : When same concurrent request comes in node then it will register an event with its callback and move ahead it will not wait for query response for a particular request.So when r1 request comes then node’s event loop (yes there is an event loop in node which serves this purpose.) register an event with its callback function and move ahead for serving r2 request and similarly register its event with its callback. Whenever any query finishes it triggers its corresponding event and execute its callback to completion without being interrupted.
So no waiting, no threading , no memory consumption – yes this is nodeway for serving rest API.

My one more reason to choose Node.js for a new project is:
Be able to do pure cloud based development
I have used Cloud9 IDE for a while and now I can't imagine without it, it covers all the development lifecycles. All you need is a browser and you can code anytime anywhere on any devices. You don't need to check in code in one Computer(like at home), then checkout in another computer(like at work place).
Of course, there maybe cloud based IDE for other languages or platforms (Cloud 9 IDE is adding supports for other languages as well), but using Cloud 9 to do Node.js developement is really a great experience for me.

One more thing node provides is the ability to create multiple v8 instanes of node using node's child process( childProcess.fork() each requiring 10mb memory as per docs) on the fly, thus not affecting the main process running the server. So offloading a background job that requires huge server load becomes a child's play and we can easily kill them as and when needed.
I've been using node a lot and in most of the apps we build, require server connections at the same time thus a heavy network traffic. Frameworks like Express.js and the new Koajs (which removed callback hell) have made working on node even more easier.

If your application mainly tethers web apis, or other io channels, give or take a user interface, node.js may be a fair pick for you, especially if you want to squeeze out the most scalability, or, if your main language in life is javascript (or javascript transpilers of sorts). If you build microservices, node.js is also okay. Node.js is also suitable for any project that is small or simple.
Its main selling point is it allows front-enders take responsibility for back-end stuff rather than the typical divide. Another justifiable selling point is if your workforce is javascript oriented to begin with.
Beyond a certain point however, you cannot scale your code without terrible hacks for forcing modularity, readability and flow control. Some people like those hacks though, especially coming from an event-driven javascript background, they seem familiar or forgivable.
In particular, when your application needs to perform synchronous flows, you start bleeding over half-baked solutions that slow you down considerably in terms of your development process. If you have computation intensive parts in your application, tread with caution picking (only) node.js. Maybe http://koajs.com/ or other novelties alleviate those originally thorny aspects, compared to when I originally used node.js or wrote this.

I can share few points where&why to use node js.
For realtime applications like chat,collaborative editing better we go with nodejs as it is event base where fire event and data to clients from server.
Simple and easy to understand as it is javascript base where most of people have idea.
Most of current web applications going towards angular js&backbone, with node it is easy to interact with client side code as both will use json data.
Lot of plugins available.
Drawbacks:-
Node will support most of databases but best is mongodb which won't support complex joins and others.
Compilation Errors...developer should handle each and every exceptions other wise if any error accord application will stop working where again we need to go and start it manually or using any automation tool.
Conclusion:-
Nodejs best to use for simple and real time applications..if you have very big business logic and complex functionality better should not use nodejs.
If you want to build an application along with chat and any collaborative functionality.. node can be used in specific parts and remain should go with your convenience technology.

Node is great for quick prototypes but I'd never use it again for anything complex.
I spent 20 years developing a relationship with a compiler and I sure miss it.
Node is especially painful for maintaining code that you haven't visited for awhile. Type info and compile time error detection are GOOD THINGS. Why throw all that out? For what? And dang, when something does go south the stack traces quite often completely useless.

We Keep Coding

JavaScript is the programming language of the Web.