10 Things to Know about Cisco UCS

WWT St. Louis Demonstration Center

Along the lines of my most read post (probably because I found guys a lot smarter than me to write it!) 10 Things to Know about the Cisco Nexus 1000V, I’ve invited Joe Onisick, WWT’s Unified Computing Practice Manager, to be my guest blogger to highlight:  “10 Things to Know about Cisco UCS.”  Here we go….and, thanks, Joe!

  1. The most important feature of UCS is its management architecture.  The hardware was all designed with unified management in mind in order to reduce the administrative overhead of today’s server environments.  As companies move to more highly virtualized environments and cloud architectures, automation and orchestration becomes key.  UCS provides the management and provisioning tools at a hardware level to quickly realize the benefits of these types of environments and maximize the inherent cost reductions.
  2. UCS is not just about blades.  The management and I/O infrastructure is designed from the ground up to manage the entire server infrastructure including rack-mount servers.  While blade adoption rates continue to grow, 60% of all servers are still rack-mount.  UCS’s ability to manage both rack-mount and blade servers under one platform is a key differentiator with major ROI benefits.  This ability will be available by the end of the calendar year.
  3. UCS is based on industry standards such as the 802 Ethernet standards and x86 hardware architecture, making it vendor neutral and fully compatible with other systems.  The UCS system is interoperable with any existing infrastructure and can be tied into management and monitoring applications already being utilized.
  4. Using the Virtual Interface Card (VIC) or Generation 1 Converged Network Adapters (CNA) from Emulex or Qlogic, UCS has a unique capability of detecting network failures and fail traffic paths in hardware on the card.  This allows network administrators to design and configure network failover end-to-end, ensuring consistent policies and bandwidth utilization.  Additionally this unique feature provides faster failover and higher redundancy than other systems.
  5. The management infrastructure of UCS is designed to allow an organization to provision and manage the system in the way that most closely fits its process.  If a more dynamic process is desired, UCS allows a single administrator to cross traditional boundaries in order to increase operational flexibility.  If the current organizational structure is rigid and changes are not desired, UCS provides tight Role Based Access Control (RBAC) tools to maintain strict boundaries that match the current customer environment.  If an organization is looking to UCS to provide an Infrastructure as a Service (IaaS) type environments, the benefits of UCS can be extended into custom self-service portals using the UCS XML interface.
  6. UCS reduces infrastructure components and costs by providing advanced tools for I/O consolidation.  The UCS system is designed to converge disparate I/O networks onto a single Ethernet infrastructure.  This consolidation is not limited to FCoE deployments; it extends these benefits to NFS, iSCSI, RDMA and any other protocol utilizing Ethernet for Layer 2 communication.
  7. Current UCS hardware provides up to 80Gbps of converged I/O to each chassis of 4-8 blades.  This is done using a pair of redundant I/O modules which both operate in an active fashion.  This is not a bandwidth limitation of the mid-plane which was designed for 40Gbps Ethernet and above.  Future I/O modules will provide additional bandwidth to the chassis and blades as data center I/O demands increase.
  8. The single-point-of-management for the server access layer provided by UCS can be extended to the VMware virtual switching infrastructure, further reducing administrative overhead.  Using Pass-Through Switching (PTS) on UCS, the VMware virtual switching environment can be managed through the UCS service profile the same way physical blades are managed.
  9. Memory extension on the UCS B250-M1 and B250-M2 blades provide industry leading 384GB of memory density for 2 socket servers.  Moreover, because this increased density is gained through additional DIMM slots, lower density DIMMS can be used at significantly lower cost to reach up to 194GB of memory.  In addition to the M250 blades, the B440 adds support for the 2 or 4 Xeon 7500 processors with 4, 6, or 8 cores depending on processor model.
  10. While the UCS architecture was designed to amplify the benefits of server virtualization and Virtual Desktop infrastructures (VDI), the platform is standards based and can be used with any bare metal x86 based operating system such as Windows, SUSE/Red Hat Linux, etc.  UCS can operate with any mix of server operating systems desired for any given customer.

Thanks again, Joe. Please be sure to visit Joe’s Blog, Define the Cloud, for more of his insight and opinions.

By the way, that’s a picture of our new demonstration center within our Executive Briefing Center in St. Louis.  We have a wide array of systems and software and can demonstrate integrated stacks from our partners including Cisco, HP, NetApp, EMC, VMware, Citrix and a whole host of other software for demonstrating systems management and orchestration.  More on our Executive Briefing Center in a future post.


8 responses to “10 Things to Know about Cisco UCS

  1. Pingback: 10 Things to Know About Cisco UCS — Define The Cloud

  2. Bob,

    As a colleague, who chooses to remain anonymous, pointed out I neglected to mention the server statlessness provided by UCS in my top ten.
    That was not done intentially, I consider statelessness to be a vital benefit of UCS due to the operational flexibility it provides when scaling up, down, performing maintenance or deploying new applications.
    To me statelessness is an inherant part of the management advantages UCS provides, but I should have mentioned it seperatley.


  3. Pingback: 10 Things to Know about Cisco UCS « Readme.txt

  4. Thanks Joe and Bob. This is a great summary of the benefits of Cisco’s innovative UCS technology and the benefits customers derive from integrating it into their data centers. One point of interest for your readers, Cisco’s B250 and C250 servers both deliver 1333MHz memory access speeds with all 48 DIMMs populated. This is a tremendous advantage for applications that are memory and speed dependent. Currently, other mainline 2 socket architectures drop to 800MHz speeds at DIMM populations over 12. This is a equates to approximately at 30% decrease in memory access performance for non Cisco UCS servers. UCS is delivering 167% more DIMM capacity without sacrificing performance, and as pointed out here, with a tremendous array of other compelling benefits as well.
    Disclosure: I work for Cisco.

  5. Ravi Balakrishnan

    Hi Joe, Bon
    Th etop 10 things to know about UCS blog makes for great reading . I want to bring to your readers additional insights on the distinct advantages offered by the Cisco VIC M81KR compared to the other Converged Network adapters out in the market

    1. Enhanced Performance
    The performance benefits of VIC come from 2 primary sources
    Elimination of software virtual switching: All VM traffic is switched by the ASIC inside the physical network switch. The hypervisor is relieved of VM switching effort and vital CPU cycles are reclaimed for workload performance.
    VMDirectPath Gen2: By associating a PCIe device directly to individual VMs, VMDirectPath Gen2 eliminates a memory copy by the hypervisor. This reduces the I/O latency while improving the throughput.
    2. Ease of use and Management
    The Cisco VIC is managed holistically by the UCSM as part of “Service Profile” associated with the server. Service profiles are created by server, network, and storage administrators using role based administration. These profiles are stored in the UCSM. Infrastructure policies, such as networking memberships, cabling requirements, server configuration and performance characteristics, are encapsulated in the service profile. Service profile templates can be used to simplify the creation of service profiles, helping ensure consistent policies within the system for a given service or application. This approach makes it just as easy to configure one server or hundreds of servers with thousands of VMs, decoupling scale from complexity. This automation reduces the number of manual steps needed, helping reduce the chance for human error, improving consistency, and reducing server and network deployment times.

    3. Technology and Business value
    CAPex: The Cisco VIC solution expands the catalog of applications that can be virtualized. Using VN-Link in HW and VMDirectPath Gen2, the Cisco VIC offers network visibility and I/O performance guarantees that were previously not possible. Demanding datacenter applications that require such guarantees can now be virtualized. This enables enterprises and service providers to further consolidate their server assets, reduce their CAPex and lower the TCO for their datacenters.

    OPex: The UCSM incorporates the management of the Cisco VIC as part of service profile. Service profiles integrate server identity, network identity and storage access into a simple, easy-to-use profile that can be used to configure hundreds of servers quickly and painlessly. Service profiles allow datacenter customers to scale without adding complexity and helps reduce the operational costs of server management and administration

  6. Ravi,

    Thanks for the reply and additional information. For clarity I’d like to point out that bullet 1 on improved performance only occurs when using the VIC alongside VM direct-path I/O. This will not be the case with non-VMware operating systems. Additionally there are limitations to direct-path I/O including no vMotion, limtits on usable NICs, etc.
    Direct-path I/O is a great feature to have available for high throughput VMs for all the reasons you state, but is not a one-size fits all solution.
    Within UCS my typical recommendation is to utilize the VIC along with UCS Pass-Through switching for ease of management. If VM counts exceed those supported by VIC virtual interfaces the Nexus 1000v is another great option for VM security, visibility and control. Additionally it puts networking back in the hands of the network team.
    For more info on the 1000v Bob has an excellent post: 10 Things to Know about the Cisco Nexus 1000v: https://bobolwig.wordpress.com/2009/08/22/10-things-to-know-about-the-cisco-nexus-1000v/
    Thanks again for the reply,


  7. Pingback: Perspective on Business & Technology: Blog’s 2010 Year In Review | Perspectives on Technology & Business

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s