Cisco Blogs


Cisco Blog > Data Center and Cloud

Simplifying Cloud Infrastructure Deployments with Cisco Common Cloud Architecture built on Cisco UCS and OpenStack

In three short years OpenStack has become cloud management platform that is “Too Big to Fail” (according to Citi Research). Whether it is true or not, OpenStack is definitely gaining traction and is making a profound impact not only as a viable Cloud management option, but also on the software economics for Cloud solutions.

Cloud computing is rapidly transforming businesses and organizations by providing access to flexible, agile, and cost-effective IT infrastructure. These elastic capabilities help accelerate the delivery of infrastructure, applications, and services with the right quality of service (QoS) to increase revenue. Cisco’s approach—innovative and unified data center infrastructure that provides the underlying foundation for OpenStack technology—enables the creation of massively scalable infrastructure that delivers on the promise of the cloud.

Cisco Common Cloud Architecture built on Cisco Unified Computing System (UCS) with OpenStack provides the foundation for flexible, elastic cloud solutions enabling speed and agility.  As the saying goes “Every Skyscraper is built on a strong foundation of pillars”, the OpenStack platform requires the core requirements from the underlying infrastructure – simplification, rapid provisioning, self-service consumption model, and elastic resource allocation. Cisco UCS uniquely provides a policy based resource management model, which simplifies by integrating compute, networking and storage with the ability to scale and automate deployment.

This foundation addresses every stage of cloud deployment be it private or public cloud offerings. Some of primary workloads targeted for OpenStack based deployments are:

  • Self-service development and test environments
  • Massively scalable software-as-a-service (SaaS) solutions
  • High-performance, scale-out storage
  • Web server, multimedia, big data, and cluster-aware applications
  • Applications with extensive computing power requirements and mixed I/O workloads

To accelerate these cloud infrastructure deployments, Cisco has developed starter configurations focused on compute-intensive, mixed or heterogeneous and storage-intensive workloads. The various server nodes are typically sized to include the OpenStack controller, compute, Ceph storage, Swift proxy and Swift storage.

 

Cisco UCS Solution Accelerator Paks for Cloud Infrastructure Deployments

Screen Shot 2013-10-31 at 10.12.12 AM

Scaling beyond 160 servers can be implemented by interconnecting multiple UCS domains using Nexus 3000/5000/6000/7000 Series switches, scalable to thousands of servers and to hundreds of petabytes storage, and managed from a single pane using UCS Central in a datacenter or distributed globally as shown in figure.

Screen Shot 2013-10-31 at 10.25.19 AM

Whitepapers:

1. Cisco UCS Solution Accelerator Paks for OpenStack Cloud Infrastructure Deployments

2. Cisco and Red Hat Simplify the Journey to the Cloud

 

Additional Information: 

www.cisco.com/go/ucs

www.cisco.com/go/openstack

www.cisco.com/go/smartplay

 

Configuration Details:

The configuration for computing-intensive workloads (including integrated management with UCS Manager):

  • Controller: 2 x C220 M3 UCS Rack-mount servers (2 x Intel Xeon E5-2665 Processors, 128GB Memory).  At least dual controller recommended for redundancy purposes. 
  • Compute/Storage: 4 x C220 M3 UCS Rack-mount servers ((2 x Intel Xeon E5-2665 Processors, 128GB Memory, 2 x 900GB SAS HDD with the option to upgrade up to 24 drives)
  • Networking: 2 x UCS 6296UP 96 10GE Port Fabric Interconnects, 2 x Nexus 2232PP 10GE Fabric Extenders.  These Fabric Extenders can support additional 10 servers, and depending upon the bandwidth requirements, these fabric interconnects in total will support up to 160 server nodes.

The configuration for mixed or heterogeneous workloads (including integrated management with UCS Manager):

  • Controller: 2 x C220 M3 UCS Rack-mount servers (2 x Intel Xeon E5-2665 Processors, 128GB Memory, 2x 600BG SAS HDD).  At least dual controller recommended for redundancy purposes.
  • Compute: 4 x C220 M3 UCS Rack-mount servers ((2 x Intel Xeon E5-2665 Processors, 128GB Memory, 2 x 600GB SAS HDD with the option to upgrade up to 24 drives)
  • Storage: 2 x C240 M3 UCS Rack-mount servers ((2 x Intel Xeon E5-2665 Processors, 256GB Memory, 12 x 900GB SAS HDD with the option to upgrade up to 24 drives)
  • Networking: 2 x UCS 6296UP 96 10GE Port Fabric Interconnects, 2 x Nexus 2232PP 10GE Fabric Extenders.  These Fabric Extenders can support additional 10 servers, and depending upon the bandwidth requirements, these fabric interconnects in total will support up to 160 server nodes.

 The configuration for storage-intensive workloads (including integrated management with UCS Manager):

  • Controller: 2 x C240 M3 UCS Rack-mount servers ((2 x Intel Xeon E5-2665 Processors, 256GB Memory, 12 x 900GB SAS HDD with the option to upgrade up to 24 drives).  Dual controller recommended for redundancy purposes.
  • Compute/Storage: 2 x C240 M3 UCS Rack-mount servers ((2 x Intel Xeon E5-2665 Processors, 256GB Memory, 12 x 900GB SAS HDD with the option to upgrade up to 24 drives)
  • Networking: 2 x UCS 6296UP 96 10GE Port Fabric Interconnects, 2 x Nexus 2232PP 10GE Fabric Extenders.  These Fabric Extenders can support additional 10 servers, and depending upon the bandwidth requirements, these fabric interconnects in total will support up to 160 server nodes.

 

Tags: , , , , , ,

Comments Are Closed