Optimize the Data Center
CiscoLive London was an incredible trip and gosh it was only 30 days ago – our first little project out of that voyage is TechWiseTV85 our latest episode on Data Center technologies. Data Center Optimization: The Next Stage is now available for your viewing pleasure in our ‘still has that new website smell’ environment we affectionately call the CVC (Cisco Virtual Connection).
This show was another exercise in self-restraint as the DC team had brought out an amazing selection – if we were hoping that a global show would mean a smaller show…we were out of luck.
So how to narrow it down?
It was not easy –
We had The Beatle on our mind and as they said ‘All Together Now’ – this became our attempt at a rally cry.
Lots to Brag About.
The Data Center guys and gals have come a long way… the Data Center Business Advantage Framework has taken root and produced a number of milestones worth mentioning:
- Over one million Nexus 10Gb Ethernet Ports now shipped
- Total number of Nexus ports in customer production: Over 7 Million
- More than 10,000 NX-OS customers
- 4,000 UCS customers
What does it all mean? Its our own double rainbow…you figure it out….what it means is that the innovation here is making sense. The standards based combination of Unified Fabric, Computing and Network Services is building a logical foundation for people to stake their future on. It’s a 650 million dollar business now growing at over 700%. The early percentages are easier to brag about now of course and its hard to imagine sustaining that pace 5 years from now…but one can dream! (more on this story)
Back to the London To Do List – In typical TechWiseTV fashion – we cherry picked a few topics and lamented the rest.
Automating the Move to Private Cloud
Mark Peterson illustrates the power of Tidal Orchestrator to make this simple.
Catalyst and/or Nexus (?)
Jimmy Ray catches up with Russel Vaughan, data center field engineer extraordinaire with a cool British accent, to get his take on how design and architecture should be approached.
The Tough Questions:
- Where do you start in Data Center Design these days?
- Is it rip and replace?
- Where do you start with greenfield design opportunities…is it a tough choice between the stalwart and valuable 6500 verses the svelte new karate kid that calls himself Nexus?
- Where’s the Interop?
More info on the FC-BB-5 standard (I/O Consolidation through FCoE)
UCS Virtual Interface Card
I caught up with Ravi Balakrishnan to better understand the role and the value proposition behind the M81KR if your want to be formal…or just UCS Virtual Interface card if your day is already running long.
Questions we addressed:
- Where does it fit?
- What kind of performance gains can be expected?
- What now goes away (if anything)?
- Where is the flexibility going to come from and is it now at the expense of productivity?
Overcoming the realities of physical I/O – one more innovation at a time…
Add this term to your vocab:
Cisco UCS Virtual Interface Card (M81KR)
DC Speed Round – So Many Demos, So Little Time…
Re-checking the list of all the DC team brought to London….what did we cover what remained on the cutting room floor?
- Scalable Layer2 Multipath Networks (check out our Fundamentals of OTV)
- Virtualized Distributed Data Center
- Efficient Global Workload Mobility
Unified Network Services
- Virtual Security Gateway (our latest Fundamentals of VSG Animation) Embedded below as well.
- ACE ANM 4.1 (the new management release)
VXI: Cloud-based Virtual Desktop Services (new demo coming soon…watch our youtube channel to get it first…_
- Virtual Interface Card (VIC) (Segment 2 with Ravi!)
- UCS Manager
- UCS Servers
- Solution Focus: HANA (High Performance Analytic Appliance)
(SAP Application using extended memory)
Network Analysis Innovation
Jimmy Ray had been waiting and waiting for the right time to play with the Network Analysis Module (NAM Blade) and finding
in London became the right confluence of events.
It’s been a love/hate relationship with Jimmy Ray and Network Management but as the technology has come along…
What all was covered?
- Monitoring and Troubleshooting (of course)
- New User Experience – pre- packaged dashboards, workflow that connect problem detection, isolation and troubleshooting,
- Techniques: Interactive Reports, contextual drill downs, One Click Packet Capture, Visual Collaboration in the Dashboards,
- A little ‘day in the life of trouble ticket’
- “Everything is slow…waaaah”
Cherry picking some notes of interest:
Response Time Monitoring from WAAS Data Sources
The NAM processes the TCP flow data exported from the WAAS and performs application response time (ART) calculations and reports. You use the NAM GUI to create a WAAS data source to monitor WAAS traffic statistics. In addition to ART, NAM monitors and reports other traffic statistics of the WAAS data sources including application, host, and conversation information.
The NAM provides different ART metrics by collecting data at different points as packets flow along their paths. The NAM provides five different collection points, each represented by a WAAS data source. Figure 2-3, “WAAS Data Sources (Data Collection Points)”, shows an example of the data collection points. The solid line represents data exported from a WAAS device and/or directly monitored traffic like SPAN.
Site Level Analysis – ‘Bucketing Data’
Cisco NAM Traffic Analyzer 5.0 introduces the capability for users to define a site, with which you can aggregate and organize performance statistics. A site is a collection of hosts (network endpoints) partitioned into views that help you monitor traffic and troubleshoot problems. A site can be defined as a set of subnets specified by an address prefix and mask, or using other criteria such as a remote device data source (for example, remote WAE device and segment information). If you want to limit the view of your network analysis data to a specific city, a specific building, or even a specific floor of a building, you can use the sites function.
You can also include multiple types of data sources in the site definition, and you can then get an aggregated view of all network traffic.
The pre-defined “Unassigned Site” makes it easy to bring up a NAM without having to configure user-defined sites. Hosts that do not belong to any user-defined site will automatically belong to the Unassigned Site.
Going back in time: (Historical Analysis)
Unlike previous versions of the NAM, in which you have to configure targeted historical reports in advance, the NAM Traffic Analyzer 5.0 stores short-term and long-term data that you can view using the new dashboards.
The NAM proactively collects and stores up to 72 hours of data at a granularity of 1, 5, or 10 minute intervals, and longer-term data with a granularity of 1 to 2 hours, This allows you to specify different time periods to view trends over time and identify potential problems.
More on all that:
The latest ‘Fundamentals of Cisco Virtual Security Gateway”