Separate REST JSON API server and client? [closed]

At Boundless, we've gone deep with option #2 and rolled it out to thousands of students. Our server is a JSON REST API (Scala + MongoDB), and all of our client code is served straight out of CloudFront (ie: www.boundless.com is just an alias for CloudFront).

Pros:

Cutting-edge/exciting

A lot of bang for your buck: API gives you basis for your own web client, mobile clients, 3rd party access, etc.

exceedingly fast site loading / page transitions

Cons:

Not SEO friendly/ready without a lot more work.

Requires top-notch web front-end folk who are ready to cope w/ the reality of a site experience that is 70% javascript and what that means.

I do think this is the future of all web-apps.

Some thoughts for the web front end folks (which is where all the new-ness/challenge is given this architecture):

CoffeeScript. Much easier to produce high-quality code.

Backbone. Great way to organize your logic, and active community.

HAMLC. Haml + CoffeeScript templates => JS.

SASS

We've built a harness for our front-end development called 'Spar' (Single Page App Rocketship) which is effectively the asset pipeline from Rails tuned for single page app development. We'll be open-sourcing within the next couple of weeks on our github page, along with a blog post explaining how to use it and overall architecture in greater detail.

UPDATE:

With respect to people's concerns with Backbone, I think they are over-rated. Backbone is far more an organizational principle than it is a deep framework. Twitter's site itself is a giant beast of Javascript covering every corner-case across millions of users & legacy browsers, while loading tweets real-time, garbage collect, display lots of multimedia, etc. Of all the 'pure' js sites I've seen, Twitter is the odd one out. There have been many impressively complicated apps delivered via JS that fare very well.

And your choice of architecture depends entirely on your goals. If you are looking for the fastest way to support multiple clients and have access to good front-end talent, investing in a standalone API is a great way to go.

---

Very well asked. +1. For sure, this is future useful reference for me. Also @Aaron and others added value to discussion. Like Ruby, this question is equally applicable to other programming environments.

I have used the first two options. First one for numerous applications and second one for my open source project Cowoop

Option 1 This one is no doubt the most popular one. But I find implementation are very much http-ish. Every API's initial code goes in dealing with request object. So API code is more than pure ruby/python/other language code.

Option 2 I always loved this. This option also implies that HTML is not runtime generated on server. This is how option 2 is different from option 3. But are build as static html using a build script. When loaded on client side these HTML would call API server as JS API client. Separation of concerns is great advantage. And very much to your liking (and mine) backend experts implement backend APIs, test them easily like usual language code without worrying about framework/ http request code. This really is not as difficult as it sounds on frontend side. Do API calls and resulting data (mostly json) is available to your client side template or MVC. Less server side processing. It means you may go for commodity hardware/ less expensive server. Easier to test layers independently, easier to generate API docs. It does have some downsides. Many developers find this over engineered and hard to understand. So chances are that architecture may get criticized. i18n/l10n is hard. Since HTML is essentially generated build time are static, one needs multiple builds per supported language (which isn't necessarily a bad thing). But even with that you may have corner cases around l10n/i18n and need to be careful.

Option 3 Backend coding in this case must be same as second option. Most points for option 2 are applicable here as well. Web pages are rendered runtime using server side templates. This makes i18n/l10n much easier with more established/accepted techniques. May be one less http call for some essential context needed for page rendering like user, language, currency etc. So server side processing is increased with rendering but possibly compensated by less http calls to API server. Now that pages are server rendered on server, frontend is now more tied with programming environment. This might not be even a consideration for many applications.

Twitter case

As I understand, Twitter might does their initial page rendering on server but for page updates it still has some API calls and client side templates to manipulate DOM. So in such case you have double templates to maintain which adds some overhead and complexity. Not everyone can afford this option, unlike Twitter.

Our project Stack

I happen to use Python. I use JsonRPC 2.0 instead of REST. I suggest REST, though I like idea of JsonRPC for various reasons. I use below libraries. Somebody considering option 2/3 might find it useful.

API Server: Python A fast web micro framework - Flask

Frontend server: Nginx

Client side MVC: Knockout.js

Other relevant tools/libs: Jquery Accounting.js for money currency Webshim : Cross browser polyfill director: Client side routing sphc: HTML generation

Jquery

Accounting.js for money currency

Webshim : Cross browser polyfill

director: Client side routing

sphc: HTML generation

My conclusion and recommendation

Option 3!.

All said, I have used option 2 successfully but now leaning towards option 3 for some simplicity. Generating static HTML pages with build script and serving them with one of ultra fast server that specialize in serving static pages is very tempting (Option 2).

---

We opted for #2 when building gaug.es. I worked on the API (ruby, sinatra, etc.) and my business partner, Steve Smith, worked on the front-end (javascript client).

Pros:

Move quickly in parallel. If I worked ahead of Steve, I could keep creating APIs for new features. If he worked ahead of me, he could fake out the API very easily and build the UI. API for free. Having open access to the data in your app is quickly becoming a standard feature. If you start with an API from the ground up, you get this for free. Clean separation. It is better to think of your app as an API with clients. Sure, the first and most important client may be a web one, but it sets you up for easily creating other clients (iPhone, Android).

Cons:

Backwards Compatibility. This is more related to an API than your direct question, but once your API is out there, you can't just break it or you break all your clients two. This doesn't mean you have to move slower, but it does mean you have to often make two things work at once. Adding on to the API or new fields is fine, but changing/removing shouldn't be done without versioning.

I can't think of anymore cons right now.

Conclusion: API + JS client is the way to go if you plan on releasing an API.

P.S. I would also recommend fully documenting your API before releasing it. The process of documenting Gaug.es API really helped us imp

http://get.gaug.es/documentation/api/

---

I prefer to go the route of #2 and #3. Mainly because #1 violates separation of concerns and intermingles all kinds of stuff. Eventually you'll find the need to have an API end point that does not have a matching HTML page/etc and you'll be up a creek with intermingled HTML and JSON endpoints in the same code base. It turns into a freaking mess, even if its MVP, you'll have to re-write it eventually because its soo messy that its not even worth salvaging.

Going with #2 or #3 allows you to completely have a API that acts the same (for the most part) regardless. This provides great flexibility. I'm not 100% sold on Backbone/ember/whatever/etc.js just yet. I think its great, but as we're seeing with twitter this is not optimal. BUT... Twitter is also a huge beast of a company and has hundreds of millions of users. So any improvement can have a huge impact to bottom line on various areas of various business units. I think there is more to the decision than speed alone and they're not letting us in on that. But thats just my opinion. However, I do not discount backbone and its competitors. These apps are great to use and are very clean and are very responsive (for the most part).

The third option has some valid allure as well. This is where I'd follow the Pareto principle (80/20 rule) and have 20% of your main markup (or vice versa) rendered on the server and then have a nice JS client (backbone/etc) run the rest of it. You may not be communicating 100% with the REST api via the JS client, but you will be doing some work if necessary to make the suer experience better.

I think this is one of those "it depends" kinds of problems and the answer is "it depends" on what you're doing, whom you're serving and what kind of experience you want them to receive. Given that I think you can decide between 2 or 3 or a hybrid of them.

---

I'm currently working on converting a huge CMS from option 1 to option 3, and it's going well. We chose to render the markup server-side because SEO is a big deal to us, and we want the sites to perform well on mobile phones.

I'm using node.js for the client's back-end and a handful of modules to help me out. I'm somewhat early in the process but the foundation is set and it's a matter of going over the data ensuring it all renders right. Here's what I'm using:

Express for the app's foundation. (https://github.com/visionmedia/express)

Request to fetch the data. (https://github.com/mikeal/request)

Underscore templates that get rendered server side. I reuse these on the client. (https://github.com/documentcloud/underscore)

UTML wraps underscore's templates to make them work with Express. (https://github.com/mikefrey/utml)

Upfront collects templates and let's you chose which get sent to the client. (https://github.com/mrDarcyMurphy/upfront)

Express Expose passes the fetched data, some modules, and templates to the front-end. (https://github.com/visionmedia/express-expose)

Backbone creates models and views on the front-end after swallowing the data that got passed along. (https://github.com/documentcloud/backbone)

That's the core of the stack. Some other modules I've found helpful:

fleck (https//github.com/trek/fleck)

moment (https//github.com/timrwood/moment)

stylus (https//github.com/LearnBoost/stylus)

smoosh (https//github.com/fat/smoosh) …though I'm looking into grunt (https//github.com/cowboy/grunt)

console trace (//github.com/LearnBoost/console-trace).

No, I'm not using coffeescript.

This option is working really well for me. The models on the back-end are non-existant because the data we get from the API is well structured and I'm passing it verbatim to the front-end. The only exception is our layout model where I add a single attribute that makes rendering smarter and lighter. I didn't use any fancy model library for that, just a function that adds what I need on initialization and returns itself.

(sorry for the weird links, I'm too much of a n00b for stack overflow to let me post that many)

---

We use the following variant of #3: Make a JSON-only REST API server. Make an HTML website server. The HTML web server is not, as in your variant, a client to the REST API server. Instead, the two are peers. Not far below the surface, there is an internal API that provides the functionality that the two servers need.

We're not aware of any precedent, so it's kind of experimental. So far (about to enter beta), it has worked out pretty well.

---

I'm usually going for the 2nd option, using Rails to build the API, and backbone for the JS stuff. You can even get an admin panel for free using ActiveAdmin. I've shipped tens of mobile apps with this kind of backend. However it heavily depends if your app is interactive or not.

I did a presentation on this approach at the last RubyDay.it: http://www.slideshare.net/matteocollina/enter-the-app-era-with-ruby-on-rails-rubyday

For the third option, in order to get responsiveness of the 2nd one, you might want to try pajax as Github does.

---

I'm about 2 months into a 3 month project which takes the second approach you've outlined here. We use a RESTful API server side with backbone.js on the front. Handlebars.js manages the templates and jQuery handles the AJAX and DOM manipulation. For older browsers and search spiders we've fallen back onto server side rendering, but we're using the same HTML templates as the Handlebars frontend using Mozilla Rhino.

We chose this approach for many different reasons but are very aware it's a little risky given it hasn't been proven on a wide scale yet. All te same, everything's going pretty smoothly so far.

So far we've just been working with one API, but in the next phase of the project we'll be working with a second API. The first is for large amounts of data, and the second acts more like a CMS via an API.

Having these two pieces of the project act completely independent of each other was a key consideration in selecting this infrastructure. If you're looking for an architecture to mashup different independent resources without any dependencies then this is approach is worth a look.

I'm afraid I'm not a Ruby guy so I can't comment on the other approaches. Sometimes it's okay to take a risk. Other times it's better to play it safe. You'll k ow yourself depending on the type of project.

Best of luck with your choice here. Keen to see what others share as well.

---

I like #3 when my website is not going to be a 100% CRUD implementation of my data. Which is yet to happen.

I prefer sinatra and will just split up the app into a few different rack apps with different purposes. I'll make an API specific rack app that will cover what I need for the API. Then perhaps a user rack app that will present my webpage. Sometimes that version will query the API if needed, but usually it just concerns itself with the html site.

I don't worry about it and just do a persistance layer query from the user side if I need it. I'm not overly concerned with creating a complete separation as they usually end up serving different purposes.

Here is a very simple example of using multiple rack apps. I added a quick jquery example in there for you to see it hitting the API app. You can see how simple it can be with sinatra and mounting multiple rack apps with different purposes.

https://github.com/dusty/multi-rack-app-app

---

Some great answers here already - I'd definitely recommend #2 or #3 - the separation is good conceptually but also in practice.

It can be hard to predict things like load and traffic patterns on an API and customers we see who serve the API independently have an easier time of provisioning and scaling. If you have to do that munged in with human web access patterns it's less easy. Also your API usage might end up scaling up a lot faster than your web client and then you can see where to direct your efforts.

Between #2 #3 it really depends on your goals - I'd agree that #2 is probably the future of webapps - but maybe you want something more straightforward if that channel is only going to be one of many!

---

For atyourservice.com.cy we are using server side rendered templates for pages especially to cover the se part. And using the API for interactions after page loads. Since our framework is MVC all controller functions are duplicated to json output and html output. Templates are clean and receive just an object. This can be transformed to js templates in seconds. We always maintain the serverside templates and just reconvert to js on request.

---

Isomorphic rendering and progressive enhancement. Which is what I think you were headed for in option three.

isomorphic rendering means using the same template to generate markup server-side as you use in the client side code. Pick a templating language with good server-side and client-side implementations. Create fully baked html for your users and send it down the wire. Use caching too.

progressive enhancement means start doing client side execution and rendering and event listening once you've got all the resources downloaded and you can determine a client capabilities. Falling back to functional client-script-less functionality wherever possible for accessibility and backwards compatibility.

Yes, of course write a standalone json api for this app functionality. But don't go so far that you write a json api for things that work fine as static html documents.

---

REST server + JavaScript-heavy client was the principle I've followed in my recent work.

REST server was implemented in node.js + Express + MongoDB (very good writing performance) + Mongoose ODM (great for modelling data, validations included) + CoffeeScript (I'd go ES2015 now instead) which worked well for me. Node.js might be relatively young compared to other possible server-side technologies, but it made it possible for me to write solid API with payments integrated.

I've used Ember.js as JavaScript framework and most of the application logic was executed in the browser. I've used SASS (SCSS specifically) for CSS pre-processing.

Ember is mature framework backed by strong community. It is very powerful framework with lots of work being done recently focused on performance, like brand new Glimmer rendering engine (inspired by React).

Ember Core Team is in process of developing FastBoot, which let's you to execute your JavaScript Ember logic on server-side (node.js specifically) and send pre-rendered HTML of your application (which would normally be run in browser) to user. It is great for SEO and user experience as he doesn't wait so long for page to be displayed.

Ember CLI is great tool that helps you to organise your code and it did well to scale with growing codebase. Ember has also it's own addon ecosystem and you can choose from variety of Ember Addons. You can easily grab Bootstrap (in my case) or Foundation and add it to your app.

Not to serve everything via Express, I've chosen to use nginx for serving images and JavaScript-heavy client. Using nginx proxy was helpful in my case:

upstream app_appName.com {
  # replace 0.0.0.0 with your IP address and 1000 with your port of node HTTP server
  server 0.0.0.0:1000;
  keepalive 8;
}

server {
  listen 80 default_server;
  listen [::]:80 default_server ipv6only=on;

  client_max_body_size 32M;

  access_log  /var/log/nginx/appName.access.log;
  error_log  /var/log/nginx/appName.error.log;

  server_name appName.com appName;

  location / {
     # frontend assets path
     root /var/www/html;
     index index.html;

     # to handle Ember routing
     try_files $uri $uri/ /index.html?/$request_uri;
  }

  location /i/ {
    alias /var/i/img/;
  }

  location /api/v1/ {
    proxy_pass  http://app_appName.com;

    proxy_next_upstream error timeout invalid_header http_500 http_502
http_503 http_504;
    proxy_redirect off;
    proxy_buffering off;
    proxy_set_header        Host            $host;
    proxy_set_header        X-Real-IP       $remote_addr;
    proxy_set_header        X-Forwarded-For $proxy_add_x_forwarded_for;
  }
}

Pro: I love the separation of API & client. Smart people say this is the way to go. Great in theory. Seems cutting-edge and exciting.

I can say it's also great in practice. Another advantage of separating REST API is that you can re-use it later for another applications. In perfect world you should be able to use the same REST API not only for webpage, but also for mobile applications if you'd decide to write one.

Con: Not much precedent. Not many examples of this done well. Public examples (twitter.com) feel sluggish & are even switching away from this approach.

Things look different now. There are lots of examples of doing REST API + many clients consuming it.

---

I decided to go for the architecture of Option #2 for Infiniforms, as it provided a great way to separate the UI from the business logic.

An advantage of this is that the API Servers can scale independently of the web servers. If you have multiple clients, then the websites will not need to scale to the same extent as the web servers, as some client swill be phone / tablet or desktop based.

This approach also gives you a good base for opening up your API to your users, especially if you use your own API to provide all of the functionality for your website.

---

A very nice question and I'm surprised as I thought this is a very common task nowadays such that I will have plenty of resources for this problem, however turned out not to be true.

My thoughts are as follows: - Create some module that have the common logic between the API controllers and HTML controllers without returning json or rendering html, and include this module in both HTML controller and API controller, then do whatever you want, so for example:

module WebAndAPICommon
    module Products

        def index
            @products = # do some logic here that will set @products variable
        end

    end
end


class ProductsController < ApplicationController
    # default products controlelr, for rendering HMTL pages 
    include WebAndAPICommon

    def index
        super
    end

end



module API
    class ProductsController
        include WebAndAPICommon

        def index
            super
            render json: @products
        end

    end
end

---

I've gone for a hybrid approach where we user Sinatra as a base, ActiveRecord / Postgress etc to serve up page routes (slim templates) expose a REST API the web-app can use. In early development stuff like populating select options is done via helpers rendering into the slim template, but as we approach production this gets swapped out for an AJAX call to a REST API as we start to care more about page-load speeds and so forth.

Stuff that's easy to render out in Slim gets handled that way, and stuff (populating forms, receiving form POST data from jQuery.Validation's submitHandler etc, is all abviously AJAX)

Testing is an issue. Right now I'm stumped trying to pass JSON data to a Rack::Test POST test.

---

I personally prefer option (3) as a solution. It's used in just about all the sites a former (household name) employer of mine has. It means that you can get some front-end devs who know all about Javascript, browser quirks and whatnot to code up your front end. They only need to know "curl xyz and you'll get some json" and off they go.

Meanwhile, your heavy-weight back-end guys can code up the Json providers. These guys don't need to think about presentation at all, and instead worry about flaky backends, timeouts, graceful error handling, database connection pools, threading, and scaling etc.

Option 3 gives you a good, solid three tier architecture. It means the stuff you spit out of the front end is SEO friendly, can be made to work with old or new browsers (and those with JS turned off), and could still be Javascript client-side templating if you want (so you could do things like handle old browsers/googlebot with static HTML, but send JS built dynamic experiences to people using the latest Chrome browser or whatever).

In all the cases I've seen Option 3, it's been a custom implementation of some PHP that isn't especially transferable between projects, let alone out in to Open Source land. I guess more recently PHP may have been replaced with Ruby/Rails, but the same sort of thing is still true.

FWIW, $current_employer could do with Option 3 in a couple of important places. I'm looking for a good Ruby framework in which to build something. I'm sure I can glue together a load of gems, but I'd prefer a single product that broadly provides a templating, 'curling', optional-authentication, optional memcache/nosql connected caching solution. There I'm failing to find anything coherent :-(

---

Building a JSON API in Rails is first class, The JSONAPI::Resources gem does the heavy lifting for a http://jsonapi.org spec'd API.

---