Cisco started to show its support for hyper convergence and software defined storage market by offering Nexus 9000 data center switches featuring 25G/50G Ethernet based on custom ASICs. The new hardware is integrated with Cisco’s CliQr application defined hybrid cloud orchestration software for deploying and managing applications across bare metal, virtualized and container environments.
The new switch is included in Cisco’s new HyperFlex systems portfolio and combines with SpringPath software with Cisco UCS servers to form a distributed storage system using solid state drives and spinning disks from federated server clusters.
Cisco’s intent with its HyperFlex systems portfolio is to provide data center admins with a single distributed, multi-tier, object based data store that converges network and compute, and offers scalability along with performance and optimization with integrated policy management.
Cisco’s HyperFlex is in lines of VMware’s EVO Rail System, and products which are being offered by other startups.
Last year it was reported that Cisco would acquire Nutanix and will release a product in support with its business partner Simplivity solutions.
But at Cisco Partnership Summit a different set of opinion was revealed by the networking giant. It said that hyper convergence solutions from start ups have first generation limitations that limit the potential of technology.
- Cisco clarified that those limitations include Inflexible and inefficient lock step scaling of hyper converged nodes;
- SDS stacks built on conventional write-in-place file systems, which hinder performance and data optimization;
- New silos of management and policies instead of integrating with existing tools.
For this reason, the company has come up with HX data Platform which tries its best to overcome these limitations by tightly melding Cisco networking, security and compliance controls throughout the application lifecycle when distributing, migrating and replicating data across storage environments.
HX Data and HyperFlex other upcoming systems also enlist the Cisco fabric network to support parallel data distribution and replication for high availability and performance, Cisco says. Data is optimized through deduplication and compression, and dynamic placement in server memory, caching, and capacity tiers not only supports redundancy but lowers storage cost.
HyperFlex system can be managed by existing UCS management and automation tools. Likewise, data services are managed through Hypervisor tools like VMware’s vCenter.
Again coming back to the newly released data center switch of Cisco i.e. Nexus 9200, the hardware is based on ASIC technology that enables customers to transition from 10/40G Ethernet switching to 10/25/40/50/100G at the price of 10/40G. Cisco says that this is possible due to the use of 16nm technology that enables 36 ports of line rate 100G on a single chip, and scales from 100Mbps to 100G on the same port.
The Nexus 9200 and 9300EX are fixed configuration top-of-rack switches that can both run in standalone NX-OS mode, while the 9300EX can also run in ACI fabric mode. Cisco also rolled out a line card for the modular Nexus 9500 chassis that supports 32 40/100G ports using the new custom ASIC, and can run in either NX-OS or ACI mode.
The Nexus 9200 and 9300EX switches are list priced at $20,000 and $22,500, respectively. The 9200 is shipping now, while the 9300EX will ship from May 2016.