Cisco released UCS Manager 2.2

Cisco UCS.jpeg

Last week Cisco released an early Christmas present, Cisco UCS Manager (code name: El Capitan), which includes a ton of new features.

For those of you who don’t know Cisco UCS Manager (UCSM), it provides unified, embedded management of all software and hardware components of the Cisco Unified Computing System (UCS) across multiple chassis, rack servers, and thousands of virtual machines. Cisco UCS Manager manages Cisco UCS as a single entity through an intuitive GUI, a command-line interface (CLI), or an XML API for comprehensive access to all Cisco UCS Manager functions.

This new release includes a ton of new features but the ones I really like are:

  • Direct Connect C-Series to FI without FEX
    • Support direct connections of C-Series rack servers to the Fabric Interconnect without having to invest in a 2232PP FEX
    • Supported for the following rack servers connected with Single Wire Management and Cisco VIC 1225 adapter: C260 M2, C460 M2, C22 M3, C24 M3, C220 M3, C240 M3, C420 M3;
  • Direct KVM Access
    • Direct KVM access launches KVM via URL: http://<IP_address of CIMC> or https://<IP_address of CIMC;
    • System admins allow server admins to access the KVM console without requiring the UCSM IP address;
    • The CIMC IP URLs are hosted on the Fabric Interconnect;
    • Supported over out-of-band only;
  • Enhanced Local Storage Monitoring
    • Enhance monitoring capabilities for local storage, providing more granular status of RAID controllers and physical/logical drive configurations and settings
    • New Out-of-Band communication channel developed between CIMC and the RAID Controller allows for near real-time monitoring of local storage without the need for host-based utilities or additional server reboot/re-acknowledgement
    • Support monitoring the progress and state of long-running operations (e.g. RAID Rebuild, Consistency Check)
  • FlexFlash (Local SD card) Support
    • UCSM provides inventory and monitoring of the FlexFlash controller and SD cards
    • Local Disk Policy contains settings to enable ‘FlexFlash RAID Reporting’
    • Number of FlexFlash SD cards is added as a qualifier for server pools
  • Flash Adapters & HDD Firmware Management
    • UCSM Firmware bundles now contain Flash Adapter firmware and Local Disks firmware.
    • UCSM Host Firmware Policies can now designate desired firmware versions for Flash Adapters and Local Disks

These features really help in minimizing VDI solution stacks because you no longer need separate FEX, like Nexus 2232, to connect rack servers to the Fabric Interconnects to manage alle UCS servers, rack or blad, with one management platform. Besides that, you can now manage local storage which you regularly need with high end VDI solutions. The direct KVM Access is ideal for shared compute environments in which you now can offer customers direct KVM access without giving them direct access to your entire management network.

Besides this, Cisco UCS Manager 2.2(1) includes the following enhancements:

Fabric Enhancements:

  • Fabric scaling
    • El Capitan supports new underlying NxOS switch code, which enables UCS to increase the scale numbers on the 6200 Fabric Interconnects, supporting up to 2000 VLANs, 2750 VIFs, 4000 IGMP Groups, 240 vHBAs, and 240 Network Adapter Endpoints.
  • IPv6 Management Support
    • Allow management of UCS Manager and UCS servers using IPv6 addresses
    • Allow access to external services (e.g. NTP, DNS) over IPv6
    • External facing client applications (e.g. scp, ftp, tftp) and external facing services (e.g. sshd, httpd, snmpd) are now accessible over IPv6 addresses
  • Uni-Directional Link Detection (UDLD) Support
    • Uni-Directional Link Detection (UDLD) is Cisco’s data link layer protocol that detects and optionally disables broken bidirectional links
    • Supported in FI End-Host and Switching mode
    • A global policy and per-port policy are added to configure UDLD parameters including: mode, msg interval, admin state, recovery action
  • User Space NIC (usNIC) for Low Latency
    • UCS will support High Performance Computing (HPC) applications through a common low-latency technology based on the usNIC capability of the Cisco VICs
    • usNIC allows latency sensitive MPI applications running on bare-metal host OSes to bypass the kernel
    • Supported for Sereno-based adapters only (VIC 1240, VIC 1280, VIC 1225)
  • Support for Virtual Machine Queue (VMQ)
    • Enables support for MS Windows VMQs on the Cisco VIC adapter
    • Allows a network adapter to dedicate a transmit and receive queue pair to a Hyper-V VM NIC
    • Improves network throughput by distributing processing of network traffic for multiple VMs among multiple CPUs
    • Reduces CPU utilization by offloading receive packet filtering to the network adapter


Operational Enhancements:

  • Two-factor Authentication for UCS Manager Logins
    • Support for strengthened UCSM authentication, requiring a generated token along with username/password to authenticate UCSM or KVM logins
    • UCSM uses single authentication request which combines (token and password) in the password field of the authentication request
  • VM-FEX for Hyper-V Management with Microsoft SCVMM
    • UCSM will support full integration with SCVMM for VM-FEX configuration
    • A Cisco provider plugin is installed in SCVMM, fetches all network definitions from UCSM and periodically polls for configuration updates
    • Supported for SCVMM 2012 SP1, Windows Hyper-V 2012 SP1 & Windows Server 2012
  • CIMC In-band Management
    • CIMC management traffic takes the same path as data traffic via the FI uplink ports
    • Separate CIMC management traffic from UCSM management traffic increases bandwidth for FI management port
    • Support In-band CIMC access over IPv4/IPv6 (IPv6 access not supported Out-of-band due to NAT limitations)
  • Server Firmware Auto Sync
    • Server Firmware gets automatically synchronized and updated to version configured in ‘Default Host Firmware Package’
    • Global policy allows user to configure options:
      • Auto Acknowledge (default)
      • User Acknowledge
      • No Action (feature turned off)
    • Guarantee server firmware consistency and compatibility when adding a new or RMA’ed server to a UCS domain


Compute Enhancements:

  • Secure Boot
    • Establish a chain of trust on the secure boot enabled platform to protect it from executing unauthorized BIOS images
    • Secure Boot utilizes the UEFI BIOS to authenticate UEFI images before executing them
    • Standard implementation based on the Trusted Computing Group (TCG) UEFI 2.3.1 specification
  • Precision Boot Order Control
    • Support creating UCSM Boot Policies with multiple instances of Boot Devices (FlexFlash, Local LUN, USB, Local/Remote vMedia, LAN, SAN, and iSCSI)
    • Provides precision and full control over the actual boot order for all devices in the system:
      • Multiple Local Boot Devices (RAID LUN/SD Card/Internal USB/External USB) and SAN
      • Local & Remote vMedia devices
      • PXE/SAN boot in multipath environments
  • Trusted Platform Module (TPM) Inventory
    • Allow access to the inventory and state of the TPM module from UCSM (without having to access the BIOS via KVM)
  • DIMM Blacklisting and Correctable Error Reporting
    • Improved accuracy at identifying “Degraded” DIMMs
    • DIMM Blacklisting will forcefully map-out a DIMM that hits an uncorrectable error during host CPU execution
    • Opt-in feature enabled through an optional Global Policy (Disabled by default)


The El Capitan features enable several UCS Solutions including:

  • VM-FEX with SCVMM for MS Private Cloud
  • Direct Connect C-Series for Smaller Big Data Clusters
  • Direct Connect C-Series for Smaller VDI Deployments
  • Direct Connect C-Series for FlexPod Reference Architecture with ESX 5.5
  • Enhanced Local Storage Monitoring for Improved System Management Integration and SMB VDI Solutions
  • PCIe Flash Cards Support for Non-Persistent VDI
  • usNIC-based HPC Solutions on Cisco UCS B-Series
  • Ubuntu Support for OpenStack


Links to download this release are as follows:

  • Infrastructure software bundle: Click here to download
  • B-series and C-series software bundles for this release are available at the above link, under “Related Software”.
  • UCS Platform Emulator 2.2(1b):  Click here to download
    • NOTE:  From UCS PE 2.2(1bPE1) onwards, UCS PE supports uploading the B-Series and C-Series server firmware bundles.  Because of the large file sizes of the firmware bundles, UCS PE only supports uploading of only the stripped-down versions (attached to this document), which includes only the firmware metadata but not the actual firmware itself in the binaries.  The stripped-down version of the firmware bundles which contain metadata only of the B-series and C-series server firmware is reduced to approximately 50 kB in size.

Cisco UCS: What’s the maximum bandwidth per blade?

Cisco UCS chassisAs I already mentioned in this article, Cisco UCS is a complex combination of components. With the release of the M3 type Cisco UCS blades the number of components have increased and so did the complexity. The result is that I receive calls from colleagues/customers on a regular basis, complaining about the amount of bandwidth per Cisco UCS blade.

I already described that, the maximum number of VIF’s depends on the number of uplinks to the chassis.

With the introduction of the M3 Cisco UCS blades we got a VIC1240 Modular Lan On Motherboard (mLOM) mezzanine adapter which gives 2 x 10Gbps to each Fabric Extender.

There’s also an additional mezzanine slot which can be used for a variety of additional hardware like a FusionIO- or LSI Nytro WarpDrive adapter. In this case we can also use this slot to expand the capacity of the VIC1240 by adding a port expander or a VIC1280. This mezzanine slot also provides 2 x 10Gbps to each Fabric Extender.

Now I learned that the amount of bandwidth for each fabric depends on the combination of the network interface(s) in the blade server and the type of Fabric Extender in the chassis.



UCS Central available NOW

Last Thursday Cisco released version 1.0 of their new multi-UCS-domain management tool,
UCS Central and now it is generally available.

UCS Central is a product that enables IT administrators to manage multiple UCS domains that are managed through individual UCS Managers from a single location through a single interface. UCS Central works closely with UCS Manager and provides consolidated management functionality for multi-UCS-installations. It uses a policy management framework similar to UCS Manager but provides a broader scope of these policies across multiple UCS domains.

In addition, it also provides consolidation of data such as hardware and firmware inventory, faults, ID pools. The product is deployed as a virtual machine that works with the multiple UCS managers. UCS Central manages thousands of physical blade and rack servers that are integrated with UCS fabric interconnects distributed across multiple UCS domains. UCS Central provides a intuitive web based GUI and an accompanying CLI for user access . It also provides an XML API similar to the one found in UCS Manager to provide automation and integration capabilities with higher level systems management tools.


Manage multiple Cisco UCS domains with one tool

VMworld 2012 San Francisco / Barcelona



Today I visited the Cisco booth here at VMworld 2012 in Barcelona and got a inside view of the UCS Central. UCS Central is a tool Cisco is working on to manage multiple Cisco UCS systems/domains. At the moment it is still in beta but they expect to release it Nov/Dec 2012.

UCS Central is serves as the UCS central nervous system and integrates all UCS components across multiple UCS domains, managing server blades, VMs, storage and networking—across multiple chassis—as a single logical domain. UCS Central provides an intuitive GUI with a command-line interface (CLI) and XML API options. It enables real-time configuration and reconfiguration of resources. The built-in XML API provides an open management framework for customers, developers, system integrators and managed service providers to interoperate with Cisco Unified Computing System resources.


10 things you should know about Cisco UCS

Two years ago the HP Vice President for Enterprise Servers Storage and Networking, Randy Seidl, said the following:

A year from now the difference will be (Cisco) UCS (Unified Compute System) is dead and we have had phenomenal market share growth in the networking space.

Man, he must feel pretty stupid right now. In Q2 of 2012 Cisco realized a 22% market share in the blade server market in North America and a 15% market share worldwide. How’s that for being dead, mister Seidl?

I have sold Cisco UCS solutions for about 18 months now and Cisco has definitely made a good impression in the server market. But there are still a lot of people (customers, (former) colleagues, VMware enthusiasts) who have vaguely heard of Cisco UCS but don’t see the distinct differences with blade server solutions like those from HP, IBM, Dell or Fujitsu.

So I decided to sum up

10 things you should know about Cisco UCS


Cisco UCS: What’s maximum number of VIFs per blade?

As one of the largest Cisco Partners in the Netherlands we do a lot of Cisco UCS implementations and as the first company in the Netherlands with the Cisco Advanced Data Center Architecture Specialization, where the place in the Netherlands for Cisco UCS troubleshooting. Last week a colleague was called to a troubleshoot a customer problem.

The customer was unable to create a 14th Virtual Network Interfaces on their Cisco UCS Virtual Interface Card and 13 interfaces is far from the maximum of 128 or 256 possible virtual interfaces per Cisco UCS VIC. Fortunately the solution appeared to be simple.

In a Cisco UCS environment all centralized intelligent occurs in the Fabric Interconnect. When using Cisco UCS Virtual Interface Cards (VICs) you can create Virtual Network Interfaces (VIFs) which can be presented to individual virtual machines. All of these virtual interfaces that are created show up in the Fabric Interconnects. They are called VIFs (Virtual Interfaces) and use VN-Tags.

The number of VIFs per blade is limited by the most restrictive item in the following list:

  • the network connectivity from chassis I/O Module (IOM) to Fabric Interconnect;
  • the Adapter VN-Tag namespace;
  • the OS/BIOS version.


Cisco UCS competitive update webinar

Cisco has been around for years and years on the networking and security side but since a year or so they are a real upcoming player on the server market. Personally I’ve had the pleasure of experiencing Cisco’s Unified Computing Systems first hand and WOW I’m impressed.

On March 23, 2011 (1:00 PM – 2:00 PM GMT) and on March 30, 2011 (2:00 PM - 3:00 PM) Cisco organizes a UCS C-Series – Competitive Update Webinar for you to learn more about the Cisco rack mount server offering.

When you attend this Webinar you will learn about UCS C-Series value proposition and receive an overview and competitive portfolio comparisons. You will also receive guidance on UCS competitive pricing and architectural comparison. Specific areas of focus are management, stateless computing, optimized virtualization, and unified I/O all of which impact TCO.

When you are also interested in Cisco rack mount servers, you can register here.