I intend to use sails.io to build a chat website. There will be some 1000 user online at the same time. is sails.io is enough to do that? And is there a way to test performance of chat website? with some normal website i known jmeter but for chat website i know nothing at all
That will depend mostly of the server you will be using for your service.
Sockets are simply an array of connections. You can have as many as you want (within normal memory usage limits of your server machine).
You can checkout this answer for more information on socket costs What's the maximum number of rooms socket.io can handle?
I am currently sails.io for a chat product with 2000+ simultaneous users using it during business hours. Sails socket.io have been holding it pretty well. Nevertheless I got it prepared for horizontal scaling when I maximum capacity starts to show symptoms.You should too.
Related
I'm a PHP developer with a lack of experience on other sever side langages.
I’d like you to give me leads, advice, keywords or whatever that could help me refine my research better.
What I want to do is basically to create a one to one mobile app chat that will scale.
There will be channels of 3 users: User A, User B and the "computer" pushing some messages according to some channels informations like the last time a message has been sent, etc.
User A should know if User B is online, writing, etc.
Every conversation should be stored in a database and would be analyzed by some algorithms. Those algorithms would also analyzed stuffs on user Facebook open graph.
The application should be able to send notification on IOS and Android.
A web administration should allow admin to set some stuff that will define what kind of message would be sent by the "computer".
I'v read lot of posts about websocket, xmpp, node.js, socket.io, etc but I don't have enough knowledge in those areas to decide what architecture should I build to make everything works together. Or maybe there is some cloud base solutions that would fit my needs...
Thanks
As you've stated there are many ways to implement that kind of structure but I am going to write about node.js + socket.io part;
1) It is scalable. You can use cluster, nginx, haproxy. etc. to apply load balancing to your socket.io application (see here) Of course you've to use redis or mongo or some kind of store for socket.io that different servers and processes can communicate each other. (see here)
2) socket.io have rooms. That means clients and any computer bots can join that room to share events with each other. So, in your scenario User A, User B and a computer bot should join to same room and events sent to that room will be broadcasted to every room member. (events can vary as online, typing, new message, anything) (see here)
3) node.js can send push notifications both for iOS and Android.
4) You can write every message to database of your choice on new message event.
5) You can create a REST api with Express framework for your Administration page. And you can use passport for authentication and authorization purposes. You can consume the rest api with Jquery or some other frontend framework like React etc.
Meteor is very well suited for something like this and gives you everything you need. There are also open sourced chat systems built with meteor already to get an idea of where you need to go. The more custom route would be to do what #cdagli said.
I have to perform load test on REST web services of 1000 concurrent users activities. I have selected NeoLoad tool for it. Is it required to have license for 1000 virtual users to simulate 1000 users or form one virtual user I can simulate 1000 users activities?
Load testing tools (like NeoLoad) price their license based on concurrent virtual users. So, if you have a license for a single user, you can create many requests, but only one at a time. To general load of 1000 users concurrently, you'll need a 1000 users license.
To simulate 1000 concurrent users you really need 1000 VUs as if you use only one you would be:
having wrongly good responsectimes due to caching effect
having wrongly negative response times due to contention on writing
you would probably not be reproducing the production behaviour of server regarding memory usage, caching....
But note that for http restful load testing you could go for a lot of free production ready open sources like Apache JMeter, locust...
https://www.blazemeter.com/blog/rest-api-testing-how-to-do-it-right
https://www.redline13.com/blog/2018/05/test-rest-apis-authentication-using-jmeter/
how can i measure stress testing for the iPhone app ?
i need stress testing not performance testing, for example 100 users access the database of the app which is on the server at the same time.
any help?
thanks in advance
First, you need to decide if you need to test the client-side (iPhone) app, the server-side code, or both.
Testing ONLY the server-side, might make this much easier - especially if it is using HTTP to communicate with the server and exchanges data via a text-based format (XML, JSON, etc). There are many web load testing tools available which can handle this scenario. Using our Load Tester product, for example, you would configure the proxy settings on your iPhone to point to our software running on a local machine. Then start a recording and use the application. Load Tester will record the messages exchanged with the server. You can then replay the scenario, en masse, to simulate many users hitting your server simultaneously. The process, at a high level, is the same with most of the web load testing tools.
Of course, the requests to the server can't be replayed exactly as recorded - they'll need to be customized to accurately simulate multiple users. How much customization is needed will depend on the kind of data being exchanged, the complexity of the scenario and the ability of the tool to automatically configure dynamic fields (and this is one area where the abilities of the tools vary greatly).
Hope that helps!
A basic simulation would involve running your unit tests on OS X, using many simultaneous unit test processes (with unique simulated users, and other variables).
If you need more 'stress', add machines - you'll likely end up hitting io or network limits from one machine relatively early on.
I would like to disregard web applications here, because to scale them horizontally, ie to use multiple server instances together, it is "sufficient" to just duplicate the server software over the machines and just use a sort of router that forwards requests to the "less busy" server machine.
But what if my server application allows users to engage together in realtime ?
If the response to the request of a certain client X depends on the context of a client Y whose connection is managed by another machine then "inter machines" communication is needed.
I'd like to know the kind of "design solutions" that people has used in such cases.
For example, the people at Facebook must have already encountered such situation when enabling the chat feature of their social app.
Thank you in advance for any advise.
One solution to achive that is to use distibuted caches like memcache (Facebook also uses that aproach).
Then all the information which is needed on all nodes is stored in that cache (and a database if it needs to be permanent) an so all nodes can access that information (with a very small latency between the nodes).
regards
You should consider some solutions that provide transparent horizontal database scalability and guarantee ACID semantics. There are many solutions that offer this at various levels. People at Facebook which you reference have solved the problem by accepting eventual consistency but your question leads me to believe that you can't accept eventual consistency.
Beside the obvious person to person instant message chat, What else have you used a Jabber server's functionality to enable?
Edit: links to working code to really show it off are particularly useful - and will be more likely to be voted up.
There are unlimited uses for XMPP/Jabber.
Take any message/data you want to send somewhere else and you can use jabber. Run a centralised logging service for distributed services? You can jabber the massage.
You want to check if your services/programs are running? XMPP presence will tell you. If you add custom status messages you can see exactly what is going on.
This is why Cisco has got into the game. Picture a server farm where each blade has a built in mini jabber client. On boot up it will register it's presence to the central server as awaiting work. The central server fires off some work in it's direction and it then changes it's status to "Busy". Another blade finished it's work and changes it's status back to "Available"... rinse and repeat.
When you combine the actual jabber messages with it's Out Of Band abilities, these servers can post where the results of the job can be found.
Anything you can think of needing to pass a message can be done with XMPP to some degree. Be this person to person, program to program, or any combination.
You could use a Jabber server to handle/broker messages between a client application and another server application.
It can actually be pretty effective.
Not me but Martin Woodward used jabber to control a "build bunny" that displays the current status of the build server.
http://www.woodwardweb.com/gadgets/000434.html
XMPP is good for sending messages back and forth between computers that don't need to be broken into chunks. They also can't be terribly big. If you use the right library, it can be pretty easy to set up.
Sending messages to a web page. Proof-of-concept: esagila.com
I plan to use it to receive notifications from my system, such as:
Process did not finish
Report was not generated on time
User needs help
I already receive many of these messages as email. But receiving an IM could be much more effective.
You might want to look at Vertebra which is...
a framework for orchestrating complex processes in a Cloud. It is designed with an emphasis on security, fault tolerance, and portability.
From the knowledge base:
Why was XMPP chosen for Vertebra?
XMPP based instant messaging can be a good alternative to search engines for information that is small, complete in itself and required frequently and repeatedly. For example, your daily horoscope - you require it daily and it is not large.
To see an example of this add astro#askme.im to your list of contacts in your jabber client (Gmail Chat/Gtalk/or any other Jabber client) and then initiate chat with this contact by sending the word "help".
Also see www.askme.im for a whole list of chat based solutions.
I've used Jabber in the past to get email notifications. Nowadays I use it for low-priority nagios notifications, it is very useful and way cheaper than SMS:
We use xmpp as both a 'bus' and a real-time API at http://superfeedr.com
Iowa State University Department of Agronomy has created this with Jabber: http://mesonet.agron.iastate.edu/iembot/
If you're a weather freak like I am, this is VERY cool stuff!
Apple implements mobileme's push service using Jabber/XMPP's subscription services to send push notifications. That is the most widespread use of Jabber for non-IM purposes I know of. This article has more details.
My friends have also built a Jabber python bot, which is kinda cute but not all that useful :-)
Edit
The most recent Next Big Thing, Google Wave, uses Jabber under the hood. Further illustrates the power of the protcol.
We have used XMPP and BOSH to enable users to communicate with a webbrowser directly and in realtime from their phone.
For example Code you can view our open source API
The vooices site also has live examples where you can control a map and play a game using your phone via your web browser: http://www.vooices.us/
I've always thought XMPP would be a good way to deliver SNMP data. OIDs are really painful, much of the system is insecure, and the SNMP traps never work quite like you want them to. With an XMPP server in the middle and a smart component to make some choices, you can use it to send out jabber or other notifications, kick off restart jobs, update web pages, or whatever else you need.
The XML data is pretty small in this case, and you can have the one XMPP server both talk to humans in message stanzas, or computers with the same protocol.