Hadoop is more pointed at data crunching. Ive been reading up on it and need at least five Piās to make that happen. A more reliable solution and probably more practical would be a couple of Piās behind one or more loadbalancer devices (hint: pfsense or the free version of Citrix Netscaler). If you use shared storage for the Blynk data you only have to maintain the server instances.
Sure, what do you want to know? It is totally not blynk related (Iām a network engineer), but I know this stufd too
A loadbalancer simply redirects your traffic to a pool of servers running some server software. E.g. Blynk, webserver, basically anything. They are used for fault tolerance as well as spreading of loads across a farm of servers.
Thanks @Lichtsignaal
For example, A raspberry can allow 1000 users to access if I connect 2 raspberries together, does it increase the number of users without compromising performance?
@tynttn I canāt see what your problem is. Piās are $10 and you will need a lot of users before your first Pi becomes overloaded. Have you actually experienced any problems or are you just expecting problems?
Why not charge the end users a nominal sum to have access to your wonderful system and then you buy as much hardware as you will ever need.