Here's a 3D quick view video of the Cisco UCS C Series systems (Rack servers) for the higher end, high performance models: UCS C250 and UCS 460.
Note: they are best viewed in full screen and HD mode
You can also access the 3D model from the Cisco Product Page.
I decided to create a few videos with 3D models of some Data Center 3.0 Architecture components. To start, the Data Center Access layer (i.e. where servers connect to), here's a quick silent video on the Cisco Nexus 5020 Unified Fabric Data Center Switch.
Note: Best seen in HD mode, and full screen
The Nexus 5000 comes in two flavors: 5010 and 5020. The Nexus 5010 has 20 embedded line-rate 10 GigE FCoE capable ports for server connectivity; the Nexus 5020 has 40 of those ports. They also have one or two expansion modules, respectively, that can be used for Fibre Channel ports to connect to a SAN.
The Nexus 5000 family is also capable of having the Nexus 2000 associated to it, to add 1 GigE ports for legacy server connectivity. That combination makes it an ideal design for datacenter access looking for a single management point, user defined oversubscription (or line-rate) access layer, and Unified Fabric support.
I decided to start playing around with Linux after so many years that I don't touch a Linux distribution. The goal is not to play just to play, but more importantly participate and contribute to the community. I also decided to document some of my random thoughts in the form of a blog post here.
Everything started with the selection of the distribution. I thought of starting with the 3 big names: Fedora, Ubuntu, OpenSuse. I used to run Redhat (when it was free, before Fedora) and Suse (before Novell). I heard Ubuntu became very popular and I did some research and found out it is the most downloaded distribution according to the dedicated distrowatch website. Debian was only for hard-core Linux users, and Mandrake disapeared somewhere in the French riviera.
After a few tests and some research, I decided to stick with the Fedora Project and become a contributor. These are some of the reasons:
Non-Technical Reasons
Oversubscription in a network is not bad, it is healthy. Best practices show that a 1:20 and a 1:4 oversubscription level is acceptable at the user campus and server access, respectively. These are based on theoretical and empiric values, that produce best practices documentations and design guides - for more information go to Cisco's Design Zone for examples on these documents.
The Data Center is getting a lot of attention when it comes to R&D and innovation lately. I'd say it started getting hot again with the new trends in cloud computing and unified "X" (i.e. Unified Fabric, Unified Compute, etc) early 2009. The latest: FabricPath technology to aggregate multiple switches and eliminate oversubscription when needed.
FabricPath is a set of capabilities in the Cisco Nexus platform that combines simplicity of Ethernet with the reliability and scalability of Layer 3 routing.
Using FabricPath you can build highly scalable Layer 2 multipath networks without Spanning Tree Protocol. When deployed across multiple Nexus chassis, FabricPath creates a flat data center switching fabric with high switching capacity, cross-sectional bandwidth, and low predictable latency.
The FabricPath Switching System improves business agility through workload flexibility and delivers operational efficiencies through network simplification, provisioning, and reduced power needs. Such networks are particularly suitable for large virtualization deployments, private clouds, and high-performance computing (HPC) environments.
Surfing through GoodReads (one of my new favorite destinations) I bumped into a great blog post by Janalyn Voigt about page margins.
It's worth reading it, and realizing that ads are filling our online lives with noise.... think about the Bing TV commercial: "Los Liiiiink!!! No!!! "
[Thanks to Sonny for sharing this with me]
Here's a crystal clear explanation of what motivates people. Here's a tip, it's not only money...
I have been updating my Dandyid lately. I also switched my mobile phone to the iphone. Those two things led me to discovering a few new Web 2.0 and social media services.
For examples, I have signed up for accounts on Goodreads.com (a social media site for logging, reviewing, and sharing books), Fring (a mobile phone federation client for voice, video, and text communications), and keep enjoying last.fm as my favorite online radio service. However, by far, my favorite entertainment has been around Yelp.com
I learned about yelp last year when I was in California and read an article in a local newspaper about it. In a nutshell, there has been a lot of controversy around this service; as a matter of fact, a year after reading that article, I still can't believe how or why it's becoming popular at all.
Yelp.com is a site for writing reviews about local businesses. They are localized (i.e. think craigslist style) and literally every business (retail, restaurant, or service) exists in the site. You can go there and check (or write!) reviews on any business in your neighborhood. The controversy is around the fact that Yelp.com reviews can significantly impact the small business owner. Be it negative reviews can send you out of business, or many positive reviews can boost sales.
The big deal is that multiple business owners have reported that after declining a paid subscription to Yelp.com's service (offered by a Yelp.com sales person to business owners) all or some positive reviews suddenly disappear and negative ones move all the way to the top. This has been denied many times by the founders of the site; however, it is impressive how many business owners in many different places claim the exact same argument. Amazing, in my opinion.
Some very nice reads are:
Two self explanatory videos. No need to comment much on them:
(1) Installing an Operating System in a Cisco Unified Computing System
Just launch the KVM, map an ISO or disk, and install the OS.
(2) Installation of ESX 4.0 in UCS.
Nothing different from installing VMware in any server platform. It is the exact same process.
Continuing with the Unified Computing System posts I wanted to touch on the concept of Service Profiles in UCS. Service Profiles is a logical construct that makes the UCS stateless. In other words, servers are logical entities that reside in a hardware platform. These logical servers can be associated and dissociated from the physical hardware with the click of a mouse. With them, you can replace, provision, and enable a server in a matter of minutes.
Quoting the configuration guide: "Each service profile serves a specific purpose: ensuring that the associated server hardware has the configuration required to support the applications it will host. The service profile maintains configuration information about the server hardware, interfaces, fabric connectivity, and server and network identity. This information is stored in a format that you can manage through Cisco UCS Manager. All service profiles are centrally managed and stored in a database on the Fabric Interconnect."
Some important characteristics of Service Profiles:
The exact and official configuration guide can be found here
I tried to create a smooth illustrative video of how to configure Service Profile and apply them to hardware. Here's my attempt at that (my excuses in advance for the going back and forth in the video):
I just wanted to show some comparison screenshots of the UCS Manager in a Windows PC vs a Mac with OS X. The UCS Manager runs on top of Java, so it is no surprise that it runs on Macintosh.
Let's see if you can tell which is which...
For all the things that I tested, everything worked in both Mac and PC Java clients. I have not had a chance to test extensively, but I expect something to not work exactly 100% in the Mac. The only detail that I found was that some windows that pop up were not sized correctly - workaround: drag the corner and resize manually.
