The SimpliVity hyperconverged infrastructure was engineered to provide a simple, scalable architecture that can deliver both predictable and peak performance while ensuring full data protection. Starting with a single node, the SimpliVity solution can be grown to a globally distributed network of nodes that can very efficiently move and protect data between physical locations, while also ensuring high availability per site. This post provides a look at the entire SimpliVity hyperconverged infrastructure from the inside of a node, all the way through to a globally-distributed, multisite implementation of the SimpliVity Federation.
The OmniStack node is the basic building block for the SimpliVity hyperconverged infrastructure. It is delivered as a 2U rack-mounted device as the SimpliVity OmniCube or integrated into a Cisco and/or Lenovo server. This choice gives customers the ability to pick the platform that is right for them based on cost, consistency with existing infrastructure components, and utilization of existing management tools.
Every OmniStack node is designed for high availability. Starting with a standard x86 server, every node includes typical redundancies like dual power supplies, fully redundant fans, multiple NIC ports, and error checking and correcting memory. The internal disk storage, both SSDs and HDDs, are protected through hardware RAID to ensure a node stays available even during multiple disk failures.
Included in every node is a SimpliVity-designed PCIe card called the OmniStack Accelerator Card. This card is connected to the OmniStack Virtual Controller utilizing hypervisor passthrough technology that allows it to avoid any latencies that may be caused by the hypervisor. The OmniStack Accelerator Card handles all data efficiency operations, providing inline deduplication and compression without any penalties and eliminating the overhead that consumes hypervisor host CPU cycles. This leaves more resources available to the business applications running on the node, while also eliminating the consumption of IOPS on the hard disk drives caused by writing duplicate data. Many customers, like Worth & Company, Inc., report significant performance improvements because of SimpliVity’s accelerated data efficiency.
Multiple OmniStack Nodes in a single site, or across two sites in a stretched cluster configuration, are combined into a SimpliVity Data Center. The SimpliVity Data Center is a logical construct within the SimpliVity hyperconverged infrastructure and is separate from, but closely aligned to, the vCenter Datacenter construct. All the OmniStack Virtual Controllers combine together to present one unified pool of resources.
Combining multiple SimpliVity Data Centers together creates a SimpliVity Federation. The Federation construct allows the utilization of backup policies across data centers, the ability to perform backups and restores between data centers, and very easily and efficiently move VMs between data centers. All sites are managed via a single interface, utilizing the native hypervisor management and orchestration tools – vCenter, OpenStack, System Center Virtual Machine Manager (coming soon!), vRealize Automation, UCS Director, etc. When adding new data centers into a SimpliVity Federation, the network topology (either hub-and-spoke or mesh) is detected and the interconnections between the different data centers in the Federation are automatically configured to match.
So there you have it, a micro-to-macro look at the components involved in a SimpliVity hyperconverged infrastructure, from the individual components within an OmniStack node to a globally distributed, multisite implementation of a SimpliVity Federation. This information and far more details of the SimpliVity platform can be found in our new SimpliVity OmniStack Technology Overview for vSphere white paper.