A look at the specs for the new Nexus UCS Invicta flash appliances. (Image: Cisco)
Seeking to address the growing focus on the role of the network in application performance, Cisco has announced an expansion of its Nexus Switching portfolio, boosting performance on Unified Computing System (UCS) compute fabrics by adding Invicta solid-state flash memory solutions based on technology it acquired with its deal to buy Whiptail last year. Cisco (CSCO) is also opening up its UCS Director to serve as a centralized control point for compute, network, and storage infrastructure.
Cisco has added new switches to match diverse workloads and use cases. Ideal for small to medium size deployments, a new 6 slot, 9U Nexus 7706 switch features up to 21 Tbps of switching capacity and supports up to 192 10GE ports, 96 40GE ports, or 48 100GE ports in a front-to-back air flow form factor. A new F3 10GE module for the Nexus 7700 chassis gives line rate 48 port, 1/10GE (SFP+), and is based on Cisco's custom F3 ASIC.
Cisco is extending the Nexus 5600 line to include the 5672UP and 56128P switches, adding VXLAN bridging and routing capability to deliver high performance in a compact form factor for 10G top of rack, deployments for data center environments. New Nexus 3172TQ top of rack 1U switch delivers 1/10G BaseT server access and performance combined with NX-OS features.
Finally, a new Nexus 1000V on the Kernel-based Virtual Machine (KVM) hypervisor brings OpenStack cloud a fully integrated network virtualization solution that can be deployed consistently across VMware, Microsoft, and Linux based platforms.
New UCS Invicta SSD models
Cisco is launching two new UCS Invicta Series systems, which leverage its its acquisition of WHIPTAIL last year. Cisco for the first time is introducing solid-state memory for UCS compute fabrics, putting it close to workloads to accelerate applications.
A 2U 3124SA UCS Invicta appliance offers the medium-scale environment a modular solution for I/O acceleration. It features 250,000 IOPS, 1.9 GB/s bandwidth, and storage capacities up to 24 TB raw or up to 64 TB with data reduction enabled. The UCS Invicta Scaling System is designed for organizations requiring enterprise-class scale, capacity, management and performance. It features up to 4 million IOPS, up to 40 GB/s bandwidth, and storage capacities up to 720 TB raw, and up to 1.9 PB with data reduction enabled. Unified Management of both the UCS Invicta Appliance and the Scaling System is done with UCS Director integration.
Cisco UCS Director Update
Cisco introduced a new version of UCS Director, a management and monitoring tool to reduce infrastructure deployment time, and manage across physical, virtual, and bare-metal environments -- supporting non-Cisco hardware (HP, IBM, and Dell) and multiple hypervisors: VMware ESX, and Microsoft Hyper-V. Evolving from the Cisco acquisition of Cloupia in 2012, UCS Director is now a leading converged infrastructure management software solution, supporting Flexpod and VCE Vblock integrated infrastructure systems.
“Provisioning a virtual machine for SAP for Mobility used to take at least two weeks,” said Pablo Lambert, Entel director of IT innovation and development. “When a customer requested a virtual machine, first a network specialist provisioned switch ports. In a day or two, someone else configured the operating system. Other administrators would install the software and middleware. Finally someone else would test the server. Now, after we’ve created a template for a certain type of server, Cisco UCS Director automatically provisions all components end-to- end. With Cisco UCS Director, we can begin using a server in 20 minutes.”
Enhancements in the new version include deeper support for Cisco UCS, Cisco Invicta solid state systems, Cisco Nexus networking, and both NetApp and EMC storage. It also ventures outside of Cisco's partnerships, and now supports HP, Dell and IBM server support, Brocade and F5 network support, and VMware and Hyper-V hypervisors. Managing truly heterogeneous data centers, UCS Director uses real-time available capacity, internal policies, and application workload requirements to optimize availability of the most beneficial or best-suited resources . It can scale to support large enterprise environments of up to 50,000 virtual machines, and 5,000 devices.
ACI Ecosystem Expansion
Cisco says its ACI (Application Centric Infrastructure) ecosystem has seen strong growth over the last year, with over 300 customers for the new Nexus 9508 switch. Cisco has added five new ACI ecosystem partners: A10 Networks, Palo Alto Networks, catbird, Cloudera, and MapR. This brings the total 31 companies that are collaborating with Cisco on open APIs, open source, open standards, and a published data model for application policy . Cloudera will bring the high performance ACI fabric to its massively scalable Hadoop distribution. APIC integration will provide policy based deployment and visibility into application performance for Big Data analysis. MapR and Cisco will collaborate to optimize Hadoop application performance on 40GbE fabric networks leveraging the ACI fabric’s dynamic load balancing, low latency, and 40G line rate performance under load to provide consistently high performance for Big Data applications. MapR and Cisco will integrate APIC policy based management to simple operations and scaling of Hadoop clusters.
