One centralized place with all developer resources. Go There
The VoltDB scale-out, distributed architecture is perfect for cloud computing. In fact, more than 50 percent of VoltDB deployments are in the cloud. VoltDB offers the high availability necessary in complex, geographically-distributed environments in one cloud-ready database.
Cloud deployments are gaining popularity for a variety of reasons: flexibility, lower (or no) CapEx, and the fact that instances can be provisioned and spun up (and down) almost anywhere. These big reasons are why 50% of our customers deploy VoltDB in the cloud.
This got us thinking about cloud platforms in general, how they perform, and which were more expensive than others. Just like in 2014, we ran the YCSB Benchmark test - this time, on four cloud platforms.
What did we discover? You can read the full report, but here are the main points:
VoltDB supports bare metal, hypervisor, and container-based cloud platforms, providing the precise deployment approach needed to maximize performance, elasticity, and support varying application framework models. VoltDB integrates with leading cloud management systems, providing database server, DBaaS, PaaS or streaming integration. Enhanced importers and exporters support PaaS integration techniques.
VoltDB is available in package formats for major cloud platforms, including IaaS clouds such as Amazon Web Services (AWS), IBM SoftLayer, Microsoft Azure, HP Helion, Google Compute Engine, OpenStack, Microsoft Server 2012, and VMware vCloud, as well as PaaS clouds such as IBM BlueMix, Red Hat Open Shift, and Pivotal Cloud Foundry. VoltDB combines the capabilities of an operational database, streaming analytics, and transactionality for public or private cloud deployment. Pre-configured versions for supported platforms are available in vendor marketplaces as well as from VoltDB.
A cloud-friendly database, VoltDB provides SQL interactivity at the speed of memory to very large datasets, limited only by cluster size. In cloud deployments, VoltDB clustering implements a shared-nothing, active/active, high-availability, fault-tolerant environment with strict SQL ACID semantics. Changes are persisted to disk at all replicas for durability/recovery and exports using VoltDB's active-active database replication capability. This is in contrast to traditional SQL and other in-memory NewSQL systems, all of which use disk-based persistence as part of computation, causing both scale-out and performance to be limited by storage. For these products, expensive solid-state disks (SSDs) or special high-performance storage nodes are often used as a solution, increasing costs and contradicting the “all software, commodity hardware” operational and cost model of Cloud. VoltDB's dynamic clustering capabilities enable easy cloud scale-out by self-adapting the cluster size and automatically re-allocating sharding domains.
VoltDB serves as a fast data front end for stream processing pipelines, easily integrating with cloud stream ingest modules including Kafka, Flume or Scribe. Using VoltDB directly instead of Storm, Samza, Dataflow, or Kinesis means developers don't need to accept reduced analytics capability or need to split streaming data into subset streams to achieve performance. VoltDB continuously exports data to Hadoop, directly to HDFS, or to leading data warehouses, enabling post-stream processing analysis.