Managing an EBS-backed Hypertable cluster on EC2 with Mesos

The Hypertable development team will be introducing the new Mesos integration and will demonstrate how to launch an EBS-backed Hypertable cluster on Amazon's EC2 with the Mesos framework. This new integration provides the ability to create and populate a Hypertable database in EC2, and then hibernate it while it is not in use. This can be a cost effective way to host data sets that are needed only periodically, since per-hour instance fees are not accumulated while the database is offline (only the EBS storage costs). Administration of the cluster is performed with a handful of simple Mesos shell commands.

We will also be reviewing the Hypertable roadmap, giving an update on the progress that has been made recently, and soliciting feedback on the feature set for the fast approaching 1.0 release.

About Mesos
Mesos (https://github.com/mesos/mesos/wiki) is a project started at U.C. Berkeley and was recently proposed (https://wiki.apache.org/incubator/MesosProposal) as an Apache Incubator project. Mesos is a cluster manager providing resource sharing and isolation across cluster applications, such as Hadoop MapReduce. When one instance of MapReduce controls a cluster, it can result in problems with isolation. It also provides a common interface for cluster programming frameworks, for example Google's Pregel and Apache Hama. It will also allow users of frameworks such as Hadoop to have multiple instances of the framework on the same cluster.