ApacheCon NA 2015

Large scale data analysis often requires merging event-based data and file-based data. Often two or more tools are required to ingest these types of data. In this presentation, Joey Echeverria will explore at how to integrate Apache Flume and Apache NiFi, which is currently undergoing incubation at the ASF, to build complex data flows that enable advanced analysis.

Apache Lens enables multi-dimensional queries in a unified way over datasets stored in multiple warehouses. Apache Lens allows queries to be executed where the data resides providing logical data Cube abstraction. In a typical enterprise multiple data warehouses co-exist, as single one does not address the needs of all workload requirements in cost-effective way. Apache Hive is one of the widely-used data warehouse in the Hadoop ecosystem. The traditional Columnar data warehouses complement Apache Hive for summarized and very frequently accessed data. Having multiple data warehouses leads to data silos that Lens aims to cut within the enterprise and provide a holistic unified access.
In this talk Sharad Agarwal and Amareshwari Sriraramadasu will present the current and upcoming features. They will also give the live demonstration of Apache Lens salient features.

Apache HTrace is a new incubator project which makes it easier to monitor and understand the performance of distributed systems. HTrace aims to provide a truly end-to-end, cluster-wide view of how requests are processed in a production distributed system-- similar to Google's Dapper or the XTrace network tracing framework.

I will talk about the architecture of HTrace and how it fits into the stack. There is a lot happening in the HTrace project, and I will discuss some of the new features that are on the horizon, such as the web interface and htraced daemon. We are working on integrating HTrace into a few Apache projects such as HDFS, HBase, and Accumulo, and we hope to have many more in the future. I will talk about how developers and users can get involved with the HTrace community.

The Apache-incubating Brooklyn project gives a deploy-and-manage framework for any application. We'll show how blueprints for standing up complex software -- on metal, in clouds, docker or paas -- are written in Brooklyn, using the YAML syntax or the Java libraries, and then how at runtime this becomes a model of the running software. This view is essentially a control plane for applications, and this talk will cover autonomic control policies including scaling, failover, and upgrades, as part of the testable, source-controlled blueprint. With example blueprints including Apache Spark, Kafka, Hadoop and Ambari, Cassandra and ActiveMQ, Brooklyn can offer many projects better integration testing, streamlined used first-touch experience, and runtime ops dashboards. And as an incubation project, we're on the hunt for more beneficiaries and contributors.

 This talk presents a case study of obscure error messages raised by Apache Subversion, based on questions raised by Subversion users who ran into them. We'll discuss what developers and users can do to help raise the quality of error messages in Subversion and other projects.