The Secret Sauce of Cisco UCS: CPU

I began my career in Cisco networking in the traditional Routing and Switching technology areas. At the time, this was one of the foremost areas of engineering expertise, prompting me to pursue the Cisco Certified Network Associate (CCNA), the Cisco Certified Network Professional (CCNP), and finally the Cisco Certified Internetwork Expert (CCIE). Technical expertise, like anything else, has a distinct lifecycle, and in order to stay in touch with the cutting edge of technology, I aggressively pursued Voice technologies.

In 2009, I was caught by surprise when Cisco announced their entry into the server market, since the proverbial landscape was already crowded by long-standing players such as IBM and HP. The new product was the Cisco Unified Computing System, which launched with the 5100 blade server chassis and 6120-XP Fabric Interconnects (pictured above). While rack-mounted equivalents (the C-series) were introduced later, the original conception was the real game-changer.

Figure 1: Cisco UCS 5100 B-Series Chassis
Figure 2: Cisco UCS 6120-XP Fabric Interconnects


Review of Server/PC Architecture

To grasp the bigger picture of what Cisco UCS offers, we need to hit the proverbial “rewind” button and understand the architecture of computers in general.  If you’ve been around the industry a long time, this might be very familiar and somewhat mundane, but if not, these hidden secrets maybe invisible to you but still important.

Figure 3: Server/PC Architecture Elements


Central Processing Unit (CPU)

While the term CPU is used freely in both technical publications and popular culture, many individuals may not fully grasp exactly what this component actually does.  Often perceived as the “brain” of the computer, this device could actually be considered the “heart” as well.  If you’ve ever tried to activate a computer without a CPU installed, you certainly realize that without it nothing happens.  Composed of numerous electronic circuits, this component performs calculations, executes instructions, and brings life to the computer in which it is installed.

The CPU has evolved substantially over the years, with new form factors, sockets, speeds, and capabilities emerging over time.  The CPU plugs into a large printed circuit board called the motherboard, using a connector called a socket.  Today, multiple sockets, and by extension, multiple processors, are commonly installed on the board itself.  To further complicate the computing equation, a processor may have multiple cores, which means that one socket can actually have multiple processors in the same physical package.


Some time ago, Bill Gates was once reportedly quoted as saying that 640K of RAM was all that anyone would ever need in a personal computer or server.  My own personal laptop today came with 4 MB of memory, a far cry from the standard of 640K.  The memory I am referring to here is Random Access Memory, sometimes referred to as RAM, DRAM, and a host of other, constantly-changing designations (I will use the general acronym RAM for convenience).

RAM also plugs into the motherboard, and interacts with the CPU and operating system to perform various functions, and store data during normal system operations.  If you’ve ever tried to remember a phone number long enough to dial it, then you have a rough idea of how this type of short-term memory works.  When the system is shut off, all contents of RAM is lost because the chips cannot hold the information without an active power source.  It’s worth noting that the amount of memory required for optimal operation continues to grow, depending on the role of the machine and the operating system being used.


Since the contents of RAM disappear when the machine is off, common sense dictates that information has to be stored in some location or everything would be a waste.  This is the purpose of storage, usually involving a series of spinning disks that store information magnetically.  Commonly referred to as “hard disks”, these devices can come in all shapes and sizes, and may exist in multiple numbers on one computing device.  Solid state drives have begun to emerge, which makes use of all circuit-based components, much like flask drives.  Larger storage arrays are available for mass storage of data as well.


Computers would amount to little more than expensive toys of they had no ability to interact with the world outside of their own internal circuitry.  In order to accomplish that worthy task, slots exist on the motherboard for specific adapters, some of which may be built into the board itself (video, keyboard, and mouse inputs and outputs).  The original adapters used a now-obsolete connector called ISA (Industry Standard Architecture), with the most current adapter type called Peripheral Component Interconnect, or PCI.  Space doesn’t permit a listing of the vast number of various adapters, but chances are your own personal experiences would help you list more than just a few.

Related Courses
Cisco Data Center Training

In this article

Join the Conversation