8 Pros and Cons of Cisco UCS
Cisco recently announced its Unified Computing System (UCS) initiative to showcase the promise and flexibility of virtualization in the data center. According to Cisco, UCS provides tighter integration with virtualization, increases IT resource flexibility and efficiency, and generates significant time and resource savings.
In order to determine if UCS is right for you, in this article we’ll explore the technology and 8 key advantages and obstacles to weigh before implementation.
What is Cisco’s Unified Computing System? UCS aims to deliver on the promise of flexible infrastructure in the data center. It combines network and server functions with virtualization support on a robust blade server system. Second, UCS integrates support for Ethernet and Fibre Channel (FC) into a unified fabric, reducing the need for components such as cards for Ethernet and SAN connectivity.
UCS Concepts
Service Profiles: Configured completely through software, the profile provides definitions of a server and its LAN and SAN network connectivity. A service profile defines a single server and its storage and networking characteristics. It operates independently from specific hardware and coupled with virtualized IOs, allows profiles to be easily reconfigured and transferred between blades.
Virtualized I/O: Traditionally, servers use many I/O adapters for Ethernet and SAN connections, and blade systems add internal Ethernet and Fibre Channel switches for aggregating these connections. In contrast, as a recent Forrester Research paper explained, “The Cisco UCS consolidates all these components by putting 10 Gb interfaces on the server blades and passing them straight through the chassis via pass-through modules. The adapters are then configured by the UCS manager to appear to the operating system as virtual Ethernet, iSCSI, or Fibre Channel over Ethernet (FCoE) adapters.” *
8 Key UCS Considerations
Now you’re thinking these details sound great, but how will UCS really function on my network? What are the costs to IT? Based on talking to Cisco experts, we have created a matrix of benefits and potential obstacles you’ll want to review before buying servers from Cisco.
| UCS Benefits | Potential UCS Obstacles |
| Virtual Benefit: With virtualization as a key underpinning of UCS, it offers greater management capabilities in mixed virtual and physical environments. More memory for server platforms, enables an increase in virtual servers hosted on each physical server. | Price: To just “kick the tires” Cisco competitors will provide a price advantage. For larger projects, Cisco is priced very competitively. |
| Greater Resource Efficiency: Flexibility of service profiles offers speedier disaster recovery and increased resource scalability | Learning Curve: Correctly applying the concept of service profiles requires significant instruction, practice, and in-the-field fine tuning. Requires extensive planning for green-field installations. |
| Fewer components: Saves significant management time, power use, and money | Locking Into Single Vendor: How does being dependent upon a single vendor for purchasing and support weigh against the technology benefits?
Note: LAN and SAN infrastructure can be supplied by any vendor’s FC or 10 Gb switch |
| Performance Advantage: Integrated and enhanced storage and network I/O provides significant performance boost, especially in large data center deployments | Existing Infrastructure Support: How does UCS infrastructure integrate and support existing Cisco devices? How does it function with your other equipment from Dell, HP, or IBM? |
Because UCS is a very comprehensive platform, here are resources that will be helpful in the consideration and purchasing process.
Understanding Service Profiles
UCS Case Study
Savvis CTO Discusses UCS Purchase Decision
Myths and Restrictions of UCS
Data Center 3.0
*”Cisco’s Big Blade Server Bet.” Galen Schreck und James Staten. Forrester Research, Inc. March 16, 2009. P.1.
Rate this article
SHARE
Metrics for UCS Success
It’s one thing to believe the UCS marketing hype, but how do you know whether performance gains have actually been achieved?
We’ll look at three common application migration scenarios (multi-tiered, e-mail, and VoIP) and how to best measure performance success before and after the UCS switch. (Note: Confirm that your monitoring solution provides visibility for applications in any virtual or physical environment.)
Multi-Tiered Applications
With multi-tiered apps, it’s critical to verify whether applications are functioning smoothly in both virtual and physical environments.
Key Metrics: Application-specific details for HTTP, database protocols (TDS/SQL), file-transfer protocols (CIFS/SMB), SMTP, and MSRPC, response times (overall and per type of request), and transaction failures
In Observer: Set up a single view of multi-tiered applications in Application Transaction Analysis.
Main Console>Trending/Analysis>Application Transaction Analysis
E-mail
The service you use will greatly affect to what extent the application can be virtualized. Specific sub-applications within MS Exchange, for example, can behave differently when virtualized, causing unexplainable performance slowdowns. So, when shifting to a UCS environment, you may want to avoid virtualizing some services. Metrics like server response time are critical for understanding whether your existing set up is overtaxed, and for quantifying improvements from the new architecture.
Key metrics: Application-specific transactions details for MS Exchange, SMTP and other mail protocols, response times
In Observer: Live statistics: Main Console>Trending/Analysis>Application Transaction Analysis
Via Trending: Main Console>Trending/Analysis>Network Trending
VoIP
Verify all VoIP protocols and call components are properly configured. Precedence settings, for example, should be set to the same setting on all parts of the call.
Key metrics: Latency, round trip delay, MOS, jitter, lost packet, packets out of order, and precedence
In Observer: Capture>Packet Capture>Start capture and press Decode>Expert Analysis tab>VoIP Events
With performance documented in pre- and post-deployment environments, your team will be able to quantify success of the implementation. Likewise, these metrics will be important for potentially identifying why your organization hasn’t seen promised gains.
Rate this article
SHARE
Share Your Worst User Horror Story
We want to hear the funniest or the worst user horror story you’ve ever experienced! The winner will have their story featured in October EYE newsletter and win a $25 iTunes gift card. Send your story.
Rate this article
SHARE
© 2010 Network Instruments, LLC. All rights reserved.
Network Instruments, Observer, GigaStor, and all associated logos are trademarks
or registered trademarks of Network Instruments, LLC.
All other trademarks, registered or unregistered, are sole property of their respective owners.
