I came across this StackOverflow post regarding the different options for realtime or near-realtime data updates in a web app, all of which are applicable to Rails:
http://stackoverflow.com/questions/11077857/what-are-long-polling-websockets-server-sent-events-sse-and-comet?rq=1 Chris On Sun, Mar 15, 2015 at 11:08 AM, Chris McCann <[email protected]> wrote: > Thanks for the suggestion of long polling. I'm definitely going to look > into that as an option. > > On Sat, Mar 14, 2015 at 6:51 PM, bradleyland <[email protected]> > wrote: > >> Our app also has a "real-time" interface. Our app is a procurement app >> that orchestrates real-time reverse auctions. Many users sit down, log in, >> and participate in an auction at the same time. There are events that >> appear to update in real time to the user, but are actually updated by >> simple polling. We considered going web sockets until we looked at what >> we'd gain and what we'd lose. >> >> Polling actually scales out really well. With a socket, your users' >> connection is always consuming a socket, which is considered a resource. >> Scaling of resources with web sockets is 1:1 with concurrent users. With >> polling, you're essentially using time division to serve (potentially) many >> more users with lower actual concurrency. App server concurrency is usually >> the first bottleneck. If your app requests are served quickly, say >> 100ms, you can serve 10 requests per second with a single app server >> process/thread. If users are happy to get updates in one second, you have >> the ability to serve 10 users with a single app sever process/thread. We >> use a polling interval of 3 seconds and the illusion of real-time is >> upheld. People simply expect things to take some time. >> >> Once you go web sockets, the server handling the web socket based >> requests must have a socket available per user. Fortunately, there are >> Rails app servers that offer better concurrency these days, but that >> concurrency can still be put to good use with polling. The other option, >> which it looks like you've already identified, is to use a PaaS provider to >> outsource that bit. As you can see by Pusher's pricing, concurrency with >> web sockets gets expensive quickly. >> >> Just food for thought. >> >> On Friday, March 13, 2015 at 8:58:13 PM UTC-4, Chris McCann wrote: >>> >>> I'm trying to put together a design for showing realtime data updates in >>> a Rails app in response to calls to an API from mobile devices. >>> >>> We recently released an Android and iOS version of our first app, Vor >>> Vision, which allows people to scan images that have an invisible code >>> embedded in them. Think "invisible QR code", only without the ugly. You >>> can check it out here: vorvision.com >>> >>> I've built a Rails backend app that hosts the API and allows a user to >>> see scans of their images in realtime. Currently I just do simple Ajax >>> polling but I want to significantly improve the app via a websockets-type >>> updating system. >>> >>> When a mobile user scans an image, the owner of that image, if they are >>> looking at the dashboard at that moment, should see the scan count for that >>> image increment, along with the geolocation of the latest scan, possibly >>> with a little highlighting or other chrome to call the user's attention to >>> the update. >>> >>> I haven't used React.js, Angular.js or any of the other client-side JS >>> frameworks, but one of these seems like a good fit for elegantly updating >>> the client side data elements. The Flux-style architecture (from Facebook) >>> seems possibly useful, if it's not overkill. >>> >>> Using server sent events (SSE) or websockets (via Pusher) seems like a >>> good fit for the server side. >>> >>> Our local Planning Center Online published this: http://developers. >>> planningcenteronline.com/2014/09/23/live-updating-rails- >>> with-react.js-and-pusher.html >>> >>> Has anyone else done this or something similar? If so, what technology >>> stack did you use? Got any pointers for me? >>> >>> Thanks all, >>> >>> Chris >>> >> -- >> -- >> SD Ruby mailing list >> [email protected] >> http://groups.google.com/group/sdruby >> --- >> You received this message because you are subscribed to a topic in the >> Google Groups "SD Ruby" group. >> To unsubscribe from this topic, visit >> https://groups.google.com/d/topic/sdruby/sBP7M1n4j1U/unsubscribe. >> To unsubscribe from this group and all its topics, send an email to >> [email protected]. >> For more options, visit https://groups.google.com/d/optout. >> > > -- -- SD Ruby mailing list [email protected] http://groups.google.com/group/sdruby --- You received this message because you are subscribed to the Google Groups "SD Ruby" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
