The All Your Base conference was a brilliant example of getting a conference right. Everything was ideal. The venue was right next to the train station and the after party was directly opposite, so staggering for the trainback home.
The network was awesome, actually getting a hard line and even a free ethernet cable in the goody bag (because who is going to remember to bring one).
And of course an amazing line up of speakers from cool companies including MongoDB (10gen), Heroku, Basho, MariaDB and many more.
Here is part one of my experiences at All Your Base.
Alvin Richards gave a great peek under the skirt of MongoDB, discussing replication, latency and how to manage data effectively in a distributed system. Some classic authors used for the JSon examples that define the mongodb “schema”. I say schema, but MongoDB does not require an up-front schema design like a relational database. All you need is a unique key and you can manage your data in what ever way you want using the flexible data document approach. This allows you to have polymorphic data, in that you can assemble data in the way that suits your application rather than being constrained by the database product.
With all this power over the data, comes great responsibility. It shouldn’t just be a “Wild West” approach. So good conversations are important between developers and database admins.
Find out more about MongoDB by getting involved in the community. And try out MongoDB with your application by using the add-on available via Heroku.
Graph databases are already used widely in the “friend-of-a-friend” networks like LinkedIn, although are generally new to the audience at the conference.
With a graph database you can have multiple connections between data creating a very intricate strcutre. So whilst the graph concept is very simple, there is much to learn about using graph database well.
In a relational model, its important to focus on the design of the data relationships. The more complexity the data relationships, the more complexity in the design of the database. The biggest hangover from complex relational database design is the join. If you traverse hundres of relationships then you are executing hundreds of JOINs.
The more joins you have to process the more your performance will be affected and the relational model computes that relationship every time you query the database.
A graph database is any storage system that provides index-free adjacency - Marko RodriguezOrientDB is an open source (Apache licence) document graph nosql dbms - supports transactions, etenended, ACID Transactions, SQL and Native Queries, Asynchronous Commands, Intents, etc.
GraphDB handels relationships as a physical Link to the record assigned when the edige is created, so traversing cost is no longer affected by db size. Eg, traversing 30milion connections in 5 seconds
Once you have a well connencted DB in the for of a Super Graph, you can cross records instead of query them. All you need is some root verticies to start traversing.
Voted by the WhiteOctober developers as the most interesting technology they wanted to see, Dale Harvey the creator of PouchDB gave a great presentation using HTML5 slides (nice). Dale currently works on Firefox OS for Mozilla and is a CouchDB contributor, so he knows what he is talking about.
Essentially, PouchDB is CouchDB for your browser. You can use it by including the pouchdb.js script in your web code with a simple
<script src=pouchdb.js> </script>
Like all flexible tools, PouchDB uses JSon as the primary data store. This gives you a great amount of flexibility in the design and use of your data.
Under the hood, PouchDB uses indexedDB when run in IE, firefox, chrome. For nodejs it uses leveldb.
Why is PouchDB here, well according to Dave - Multi Master replication is awesome
If you have an app and you want to work offline then this PouchDB is a great idea. I wish something like this was part of GoogleDocs so I could continue to edit documents when I loose connection to the Internet. There are also many other mobile use cases where you need the data locally, not just because of limited connectivity but also for speed.
jr0cket Offline is still a fact - people still need some data locally especially when on the move
So what’s wrong with building an active sync into your app? Well it is very challenging. Sync’ing is very hard, especially when it comes to conflict resolution. PouchDB provides a very simple mechanism to handle it. Every document is referenced by a single id, if you change the document name then is still knows its the same document. Only when you change the id of the document does it become a different doc.
Different changes are saved separately so it allow you to merge conflicts through the PouchDB API.
This sounds like something I’d like to include in mobile projects, so hope to get some time to play with PouchDB over the holidays.
The MySQL project started from a desire to help people and do lots of travelling. After MySQL was aquired by Sun (and before aquired by Oracle) new life was breathed into the open source project. The original creators behind MySQL shifted to start working on MariaDB, a community developed database with more features yet still compatible with MySQL.
There is no reason to use MySQL anymore, MariaDB is better in every respect. Its completely open and all the closed source features added to MySQL has been added to MariaDB already. It seems that Oracle has lost much of the engineering talent around MySQL and is not able to do much around the product, hence the closed source add-ons.
One goal of MariaDB is to make a bridge between Relational and NoSQL models. Using a concept called Dynamic Columns, a proof of concept of this was done by creating a storage engine for Cassandra.
When you want to make MySQL obsolute (which is what Monty does) then you have to provide something better to replace it. In all respects it seems that is what the MariaDB project has achieved. From additional in features, major performance improvements, community support and documentation then MariaDB has already become a valuable successor.
@montywi Try MariaDB for yourself, it will take you less than 15 seconds to migrate, depending on how fast you type.
Break for lunch… see you soon.
This work is licensed under a Creative Commons Attribution 4.0 ShareAlike License, including custom images & stylesheets. Permissions beyond the scope of this license may be available at @jr0cket