NoSQL Benchmarks NoSQL use cases NoSQL Videos NoSQL Hybrid Solutions NoSQL Presentations Big Data Hadoop MapReduce Pig Hive Flume Oozie Sqoop HDFS ZooKeeper Cascading Cascalog BigTable Cassandra HBase Hypertable Couchbase CouchDB MongoDB OrientDB RavenDB Jackrabbit Terrastore Amazon DynamoDB Redis Riak Project Voldemort Tokyo Cabinet Kyoto Cabinet memcached Amazon SimpleDB Datomic MemcacheDB M/DB GT.M Amazon Dynamo Dynomite Mnesia Yahoo! PNUTS/Sherpa Neo4j InfoGrid Sones GraphDB InfiniteGraph AllegroGraph MarkLogic Clustrix CouchDB Case Studies MongoDB Case Studies NoSQL at Adobe NoSQL at Facebook NoSQL at Twitter



NoSQL Brazil Recap

Last weekend, Sao Paulo, Brazil ☞hosted the first NoSQL event. While I’d love to be part of all these events and report live from them, this is not really possible right now. But myNoSQL is extending his coverage with the help of the official ambassadors and here is Gleicon Moraes’ report from the no:sql(br).

This is the NoSQL Brazil (nosqlbr) recap. The conference was great, the room was full most of the time and the technical level was constant between presenters and audience questioning. Permeating the event was the feeling of using the right tool for the job, and that was great. Seemed like a consensus that noSQL is about not only SQL and everyone was very compromised with it. Its a bold statement for Brazil to have such a successful conference about something that even outside here is a new ground. Better yet is to know that speakers and audience alike were very well informed and experienced.

First there was the Opening, with the organizer - Alexandre Porcelli, lead developer for ☞ OpenSpotLight, which ranged from the event motivation and history (it jumped from a small meeting to a full day event at a hotel room), the motivation behind NoSQL and his involvement with it, from the beginning to OpenSpotLight.

Then I presented my noSQL and SQL anti-patterns talk - which was about my motivation to go noSQL, some SQL anti patterns I found (and was responsible for). The talk went well and I had a great time. The audience was very technical and I think some of my feelings hit home with them.

Luis Fernando Teston from OpenSpotLight brought a video from Salvatore Sanfilippo and gave a workshop on Redis, from compilation to data types and patterns. He discussed key/value datastores and gave some tips on good key names.

Guilherme Silveira from ☞ Restfulie fame, talked about REST and its meaning in HTTP enabled noSQL. He explained the hows and whys of being really HTTP compliant, from headers to caching politics. Also, he showed how he implemented it in a thin layer over CouchDB using Sinatra. His presentation style is energetic and I think he make it clear why self discovery and proper HTTP handling are healthy for such kind of databases.

Straight from 10Gen, Alberto Lerner gave a solid MongoDB introduction and explained how things works in 10Gen and MongoDB’s community. He talked about data management as a mission which must guide about the right technology and stressed out that the brazilian developers must take part in the community, ask questions, send code and take part in the projects they use. It was really graceful from the 10gen folks to send Alberto down here.

After lunch, Rodrigo Strauss presented Tio - the first brazilian noSQL. He explained his motivations and architecture from ground up. Its core is a pub/sub engine and there is a lot of use cases. Its based on the proven Boost library. Besides Tio, there is Primo and Tia and clients written on C++ and Python. He seems to be developing it alone, so any community help is welcome. The URL is ☞

RedHat guys Edgar Silva and Samuel Tauil presented their enterprise products and solutions for virtualization and JBoss based RestEasy.

Vinicius Carvalho from ☞ Sambatech brought a Hadoop showdown on his experience using it in a proof-of-concept distributed log processing system which he is using to extract useful info from webserver logs. He explained the big picture on hadoop and pig, and the best route for the first contact, which is using a hosted Hadoop service. Along with that he gave some hints on the size of the data he is managing from Sambatech services. That relates to noSQL because one of the ways to keep a constant flow of useful information coming for a database - any kind of it - is using Map/Reduce. Also, most of the databases implement some kind of internal Map/Reduce querying system.

Closing up the talks, Julio Viegas from Globalcode/SPC (local consumer protection service) showcased a very creative way to introduce noSQL on organizations and a very careful study on different types of database servers. He choose cassandra for a 3TB/5TB cache system, which runs along with a big Oracle installation. His setup has 10 servers and there was a rundown on important configuration tips along with a practical explanation on the writing/reading constraints.

After that there was an informal chat between presenters and audience about how to introduce these new tools to organizations and how to show them that the community makes non relational databases more trustable.

Even in the breaks there were groups talking about very technical issues and real life systems that use or could benefit of other data storage and management paradigm. Kudos to Alexandre Porcelli which in a ultra short time span organized and make that conference happen. Thanks to all who donated money and time (there still some bills to pay) but I wondered how much he could do with proper funding.

Pictures, more info and presentations will be avaliable later this week at ☞

Gleicon Moraes is a developer, hopefully specialized on distributed and large data systems. Had worked on major brazilian internet providers in the last 14 years and wrote a book about linux and programming in 2005 after a lot of articles for local magazines and one for Sysadmin magazine. Nowadays I keep a blog on my tech stuff ☞, projects on github: ☞ My interest on noSQL is having the right tool for the job, which is to handle a lot of data generated by my startup.