So in other words, they misused MongoDB and because of that are telling people not to use it? Wow. Seems to be a case of "a bad mechanic will always blame his tools".
In the right hands MongoDB can be a great asset. The problem here is that Diaspora chose MongoDB when it was very immature and it seems the choice was based on hype more so that mapped out requirements. This is where proper planning for a large scale application will spot these kinds of problems before they get to the development stage.
Later versions of MongoDB are much better and the upcoming planned changes will take it many steps in the right direction towards being a viable alternative to a traditional RDBMS. Having said that, it's not a silver bullet and MongoDB is not for everything.
10Gen are exceptionally great at marketing Mongo and I get the feeling they have kind of trapped Foursquare who have been using it in production for a couple of years or so now. Having said that with exception of that one 11 hour outage Foursquare encountered, MongoDB seems to be working really well for them and they seem to be capturing more than just check-in data.
I still think with proper planning pulling off a social network with MongoDB is possible. I am currently building a social networking type application, not on the scale of Facebook but it does share some parallels. I've planned and mapped out a viable structure and how it all connects, prototyping and testing seems to indicate that MongoDB is up to the task, but we'll see.
"So in other words, they misused MongoDB and because of that are telling people not to use it? Wow. ... The problem here is that Diaspora chose MongoDB when it was very immature"
So the specific misuse of MongoDB was to use it? I guess that was also the point of the story.
Yes, exactly. They chose the wrong tool for the job, that's not a fault with MongoDB. It's like eating a soup with a fork, you'll eventually eat the soup, but if you used a spoon in the first place there would never have been a problem. Is the company who made the soup to blame or are you to blame for not consuming it correctly?
The problem with Diaspora was that it was a poorly executed good idea. Had they actually sat down and mapped out their requirements and chosen an appropriate database they would have realised that a traditional relationship was the right choice to make.
Databases are hard, for a project as ambitious as Diaspora proper planning is key and evident by the issues that Diaspora had when it debuted (delete controller with no auth checking...) it's apparent they had no clue what they were doing not from just a database planning point-of-view but a code one as well.
I would take whatever anyone who had anything to do with Diaspora had to say with a grain of salt. The real lesson here is to not get caught up in hype and trends. NoSQL isn't a magical solution that will make scaling problems disappear, traditional databases like MySQL have been battle-tested over a very long period of time. If MongoDB and <insert X NoSQL database here> were so great, the likes of Facebook and whatnot would use them on the scale Diaspora tried to use them on.
People have to realise when Diaspora used MongoDB it was a very early version of the database. A lot has changed since Diaspora used it, it is suitable for a large-scale Facebook clone as the sole database? Definitely not, but using it for aspects like messaging and notifications I think would definitely be a good use case for it.
If MongoDB was the wrong tool because it was still too immature, i.e. broken, wasn't the problem not that they were eating soup with a fork, but with a broken spoon.
But yes, they were wrong to use a broken tool for the job.
In the right hands MongoDB can be a great asset. The problem here is that Diaspora chose MongoDB when it was very immature and it seems the choice was based on hype more so that mapped out requirements. This is where proper planning for a large scale application will spot these kinds of problems before they get to the development stage.
Later versions of MongoDB are much better and the upcoming planned changes will take it many steps in the right direction towards being a viable alternative to a traditional RDBMS. Having said that, it's not a silver bullet and MongoDB is not for everything.
10Gen are exceptionally great at marketing Mongo and I get the feeling they have kind of trapped Foursquare who have been using it in production for a couple of years or so now. Having said that with exception of that one 11 hour outage Foursquare encountered, MongoDB seems to be working really well for them and they seem to be capturing more than just check-in data.
I still think with proper planning pulling off a social network with MongoDB is possible. I am currently building a social networking type application, not on the scale of Facebook but it does share some parallels. I've planned and mapped out a viable structure and how it all connects, prototyping and testing seems to indicate that MongoDB is up to the task, but we'll see.