Converged & Hyperconverged Infrastructure

 View Only

Hitachi Virtual Storage Platform (VSP) Storage Capabilities with Cisco Unified Compute System (UCS)

By Arvin Jami posted 09-03-2024 13:05

  

In an era where businesses are defined by the services and capabilities provided to end customers, choosing the best in-class storage system in conjunction with a premier compute system is key to business success. With multiple vendors, it becomes more difficult to gauge what features and capabilities would actually meet the target ROI to not only sustain but grow their business.

Cisco Validated Designs (CVDs) consist of systems and solutions that have been designed, tested, and documented to facilitate and improve customer deployments. That is where the experience of Hitachi and Cisco join to provide the Cisco and Hitachi Adaptive Solution (CHAS) which combines the best-in-class storage in the form of Hitachi Virtual Storage Platform (VSP) and Cisco Unified Compute System (UCS) in an optimized, pre-validated, and tested architecture. CVDs for CHAS are extensively tested and verified within Cisco labs and enable customers to hit the ground running without surprises during initial deployment.

Each validated architecture provides customers the flexibility of choosing their configuration based on business needs. Hitachi VSP can provide an array (no pun intended) of storage delivery options, including Fibre Channel (FC). Both FC-SCSI and FC-NVMe configurations and best practices of these protocols have been documented within the CVDs, and offer redundancy in the form of multiple fabrics providing a highly available environment resilient to down time. The following figure represents the 32 Gbs end-to-end data path using the 5th Generation Cisco UCS Virtual Interface Card (VIC) 15231, the 5th Generation Cisco UCS 6536 Fabric Interconnect, and the Cisco UCS X9108-100G Intelligent Fabric Module to deliver 32 Gbps of Fibre Channel (SCSI and NVMe) connectivity from the VSP 5600 storage system to the Cisco UCS servers.

This multi-fabric approach includes several Hitachi VSP Fibre Channel ports that connect into the Cisco MDS A or B fabrics which can in turn be used for providing disk for boot LUNs as well as application volumes. For more information on this data path see the following YouTube video:

https://www.youtube.com/watch?v=r8LPIZvmIp0

The forementioned data path provides the basis for deploying hypervisor-based applications that run on VMware ESXi. Within the Cisco and Hitachi Adaptive Solutions with Cisco UCS X-Series, VMware 8U1, and Hitachi VSP 5600 CVD document, readers are guided to stand up a greenfield UCS ecosystem backed by the Hitachi VSP 5000 series which provides Fibre Channel capabilities that also include VMware APIs for Storage Awareness (VASA) vVols to satisfy the needs of businesses that are looking for not only a flexible system, but a robust enterprise storage system that can provide optimal performance and ROI.  The following figure shows the topology of the CHAS.

To expand and build on this architecture, the Cisco and Hitachi Adaptive Solutions for Epic Workloads, Design and Benchmark Guide presents the VSI architecture as a ready-made environment for Epic workloads. Cisco and Hitachi followed Epic best practices to run their IO simulator GenerateIO (GenIO) on top of VMware ESXi. Epic developed their GenIO simulator to simulate workload scenarios that are commonly seen in real-world Epic production environments to ensure that a storage solution meets the critical service-level objectives. Hitachi introduced a new feature to the VSP family called Thin Image Advanced (HTI Advanced) that builds on top of legacy Thin Image (HTI) technology. Older systems that used HTI would have higher load on storage system performance because a data copy was required every time data was written by the host. With HTI Advanced “redirect on write” snapshot technology is used to help customers solve the performance drop issue in PSUS (split) state in an Adaptive Data Reduction (ADR) setup since the write data from the host is not overwritten when the pair is split and is written to a different location within the backing dynamic provision pool, negating the need for data copy. This feature provided great results in qualifying the CHAS for Epic workloads.

Additionally, this architecture can utilize various hardware components identified by Cisco’s Hardware Compatibility Listing and Hitachi’s Product Compatibility Guide.  With the Cisco and Hitachi Adaptive Solution VDI for VMware Horizon 8 VMware vSphere 8.0 U2, this common architecture was validated and tested to support the Virtual Desktop Infrastructure (VDI) backed by Hitachi VSP E1090, a powerful mid-range system that provides the same feature set and capabilities as the VSP 5000 series thanks to the common Hitachi Storage Virtualization Operating System (SVOS).  The VSP E1090 which is ranked number 1 in Energy Star Midrange Storage Platforms provided excellent performance on LoginVSI benchmarks, while providing exceptional data reduction and capacity savings numbers which directly translate into reducing capital expenditures and maintaining a high ROI.

For enterprise customers that require high availability (HA) and 100% uptime, Cisco and Hitachi Adaptive Solutions for Converged Infrastructure as a Stretched Data Center is a best practice data center architecture extended between locations as a seamless environment for the underlying Cisco network, Hitachi storage, and VMware hypervisor-based compute infrastructure.  This architecture expands the Cisco and Hitachi Adaptive Solutions for Converged Infrastructure with Cisco ACI utilizing the Cisco ACI Multi-Pod design for a uniform network and incorporates Hitachi global-active device for storage resiliency between locations and complies with the VMware vSphere Metro Storage Cluster (vMSC) specifications. Global-active device implements cross-mirrored storage volumes between two Hitachi VSP systems accepting read/write I/Os on both sides that are continuously updated. If a disk controller failure occurs at one site, the controller at the other site automatically takes over and accepts read/write I/Os. This enables production workloads on both systems, while maintaining full data consistency and protection. Global-active device assures that an active and up-to-date storage volume is available to a production application despite the loss of a virtualized controller, system, or site.

For a full list of Cisco and Hitachi CVDs check out Cisco Validated Design Zone for Data Center.

With the next generation of Hitachi storage hitting the market known as Hitachi VSP Block One, we look forward to future CVDs utilizing 100 Gb NVMe/TCP IP for providing application support that builds on top of Hitachi SVOS and current protocol capabilities. Furthermore, the latest Cisco UCS X-Direct modules will also be incorporated which provides a smaller upfront cost and footprint when compared to traditional CHAS deployments.

 

Best,

 

Arvin

0 comments
10 views

Permalink