Instructions to download, install, configure and run BigFoot Analytics as a Service in your private cloud.
In what follows, we will guide you to download configure and use BigFoot Analytics-as-a-Service. BigFoot has been designed for private clouds, running the Apache OpenStack orchestration framework, also often called the cloud operating system. BigFoot extends OpenStack with the Analytics-as-a-Service paradigm, and allows the creation and execution of BigFoot services with an easy Web interface.
For more information about the BigFoot architecture, we prepared an high-level description of its philosophy, its components and how they interact. You can find this information in the High-level BigFoot architecture. document.
The Analytics-as-a-Service components of BigFoot have been contributed upstream to the Apache OpenStack project. As a consequence, you guessed it right! You first need to get OpenStack up and running in your own private data-center. In this section, we provide instructions and links to the authoritative Web Site with a step-by-step guide on how to download and install Apache OpenStack.
When relevant, we also provide additional documentation to help with OpenStack configuration, which can be very cumbersome for the novice! This documentation was delivered as part of the work in the BigFoot project.
We assume you now have a brand new installation of Apache OpenStack, and its companion project OpenStack Sahara, which enables the Analytics-as-a-Service experience! In what follows we will show you how to populate your private cloud with BigFoot Analytics Services. A little background first: BigFoot Analytics Services come packaged into Virtual Machine images, and we have many flavors, depending on your application use cases!
For example, you could be a complete newbie with parallel processing systems: you have data, you heard that there is this cool new framework called [Apache Spark] and you want to give it a try! In this case, BigFoot -- through OpenStack Sahara -- helps you in hassle free deployments of Apache Spark. As an administrator, a few steps are sufficient to create clusters. As a user, spinning up an Apache Spark cluster is as easy as one click! In addition, BigFoot brings you the true Analytics-as-a-Service experience: users do not even need to know about clusters, they only need to say where the data is, submit the analytic queries, and sit and relax... waiting for the results to come. BigFoot services also feature cutting-edge analytics experience, by exposing Spark Notebooks!
Instead, you could be a data science ninja, and want to spend less time in exploring and understanding your data, and more time having fun! In this case, you may want to use this cool new BigFoot service that we called DiNoDB. DiNoDB allows you to hook on your favorite HDFS data store, and start querying for your data using plain SQL! We are very excited about the DiNoDB service: learn more by giving a look at [Our Storage Services documents]
OK, enough with words... show us a step-by-step guide. In what follows we give two examples based on pre-built virtual machine images. If you want to create custom made images, please contact us to learn more.
A note on the VM image we will use: this image includes Apache Spark and the Hadoop Distributed File System (HDFS); thus, it is lightweight, but it doesn't contain other BigFoot services.
This VM image, includes DiNoDB, Apache Spark and HDFS. In addition, a client interface to query HDFS data, using the familiar Postgres SQL syntax, will also be installed on a client-facing virtual machine. The steps are the same as above, just repeat with us:
As a user, once your administrator has created templates for your organization, deploying a cluster with BigFoot Analytics Services is an easy process.