Archive for April, 2011

I’d Rather Fight Than Switch

Posted in Cisco Certification with tags , , , , , , , , , , , , , , , , , , , , , , , , , , , , , on April 26, 2011 by jjrinehart

My maternal grandfather smoked Tareyton cigarettes, which carried the famous ad line of “I’d rather fight than switch” that I thought was a perfect lead-in for discussing LAN switching (clever, right?)  To the uninitiated, the term switch conjures up a whole set of images, usually relating to electrical components and/or lights.  In terms of networking, switching was a quantum leap forward, especially in terms of network congestion and bandwidth.  Think of a hub just like a parking lot after a sporting event–there is only one exit, and everybody is crowding to get out of that one opening, and coming from all directions as well.  Not a pretty picture, especially if you have waited seemingly hours to get out…

Wouldn’t it be amazing if there were a whole block of separate exit points from that parking lot?  Imagine how much faster things would go and how much more smoothly would the lot empty.  That’s the essential idea behind switches…instead of all stations sharing a single network entry point, separate data channels are created for all of the attached devices.  Sounds almost magical, but there is a rather simple logic that makes it all work, and it all based at Layer 2…the data-link layer of the OSI model.  Hubs just took electrical signals and retransmitted it out all ports because it had no way of distinguishing traffic.  Switches are different because they are examining the MAC (hardware) addresses of the frames passing through their ports.  If the frame is one it has never seen before, or a broadcast (ffff.ffff.ffff), it sends it out all ports except the one it arrived on (termed flooding).  The switch then records the source address contained in the frame, as well as the interface it came in on, and when a frame destined for that address arrives again, it send it out the port contained in the table entry (termed forwarding).  If it arrives on the same port in the table, it simply drops the frame (termed filtering).  The reason for the term switch is simple, because it takes a frame from one port,and then switches it to another port and sends it on its way.

In the next blog, I will drill into the specific types of switches in the Cisco product line and how they are best used in real-world settings.

– Joe

Hubs, Bridges, and Switches Oh My!

Posted in Cisco Certification with tags , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , on April 14, 2011 by jjrinehart

Any of you that recognize that modified quote should be eligible for a free latte or something, and yes it’s ridiculously corny.  So am I, and my humor is rather twisted as well…

As I mentioned previously, in the LAN environment Ethernet won out over all of the other competing technologies because of its lower cost and higher bandwidth capabilities.  One weakness was the distance limitation, 100 meters or 382 feet, as well as the desire to attach multiple workstations to allow for more resource distribution.  Necessity is the mother of invention, so repeaters and hubs were introduced.  The main difference between a hub and a repeater is the number of ports: A repeater only has two (to extend distance), and a hub has multiple.  At the time, this was a major leap forward, but in modern networks they are considered obsolete, for reasons we will discuss later.  The mechanics of these devices is ridiculously simple, they receive electrical signals (which by definition is a layer one function) and retransmit the exact same signals out every port.  One drawback is that because it operates that way, bandwidth is shared, with each user diminishing the available traffic, as well as a massive increase in collisions; think of hubs as a parrot: they only repeat what thy are told…what they are told…

As networks increased in size, this reached a saturation point that had to be addressed, leading to the development of bridges and switches, which as with repeaters an hubs, can be distinguished by how many ports are involved (2 for  bridge, multiple for  switch).  Switches operate at Layer 2, not simply repeating electrical signals, but build tables based on hardware addresses (Media Access Control or MAC).  Switches are more complex and deserve a deeper discussion in a later blog.

More to come..

Joe

Cabling: It Ain’t Sexy, But It’s Got Teeth!

Posted in Uncategorized with tags , , , , , , , , , , , , , , , , , , , , , , , , , , , , , on April 13, 2011 by jjrinehart

In case you missed it, this is a quote from the movie The Firm.

When last we saw our heroes…oops, wrong channel, makes it sound more like the Rocky and Bullwinkle Show (LOL)!

Cabling in Ethernet networks is another topic that eager student engineers can skip over, and miss out over the many lessons to be had.  In thinking back to the review of the OSI model, remember this: when troubleshooting always start with the physical layer!  In my years in the industry, I have banged my head over a problem only to discover it to be caused by a cabling issue.  On a side not, if you suspect this kind of problem, have a known good cable handy to swap out as a first step, you would be amazed at how often that helps.  In any case, there are two basic copper cables used in Ethernet LANs:

Straight Through Cable

1. Straight-Through Cable: This term describes the vast majority of Ethernet cables and patch cords (shorter cables used from patch panels to devices in a communications closet).  The “pins” on an RJ-45 jack run straight through without alteration to the pins on the jack on the other end of the cable.  Straight-through cables connect switches to workstations, IP phones, wireless access points, and routers (essentially “computers”).  Computer-to-computer, and router-to-router, an switch-to-switch connections, hosever, require a different cable.

Crossover Cable

2. Crossover Cable: Crossover cables are used in specialized situations, and used far less frequently.  These cables differ from their “cousins” in that all of the RJ-45 pins do not run straight through, but rather, the transmit and receive pairs are crossed (hence the term crossover). For router-to-router, switch-to-switch, and computer-to-computer connections, this is the cable to use.

The Exception:  As with most English grammar rules, there is an exception to this, sort of.  Most of the newer Cisco switches support a feature called Auto-MDX, meaning that the hardware can support automatic crossover when needed.  With this functionality, straight through cables are all that is needed.  That being said, I have personally encountered situations where an actual crossover cable is all that actually worked.  The moral of the story?  Know how things work, and be prepared for contingencies.  Next up will be about switches and hubs…

– Joe

Local Area Network (LAN) Fundamentals

Posted in Cisco Certification with tags , , , , , , , , , , , , , , , , , , , , , , , on April 11, 2011 by jjrinehart

One of the most disruptive inventions of the previous century was the introduction of the personal computer by IBM.  IBM had already made a big splash with the introduction of the mainframe, which dominated the landscape for years.  The problem was, however, that very few homes were likely to install this huge machine and make regular use of it.  The story goes that the very first IBM Personal Computer was designed and built from readily available, off the shelf parts, and assembled rapidly.  Needless to say, it did become a huge hit, and proliferated the use of computers in homes and businesses.

Early on, business user wanted to share information, and this first happened by exchanging floppy disks, and because it usually involved walking to another cubicle, it came to be known as sneakernet.  Eventually, more technical attempts yielded the Local Area Network (LAN) which enabled users to share printers, files, Internet connections, and so forth.  The following is a simplified representation of a LAN:

Local Area Network

In the early days there were several different competing LAN technologies, such as Ethernet (spoken without a lisp), Token Ring (one ring to rule them all), ARCnet (no, not Noah), and Fiber Distributed Data Interface (FDDI, pronounced “fiddy’).  Ethernet won out because it was cheaper, had better bandwidth, and wider option, which created a lot of new standards.  Speeds and feeds begin at 10 Mbps and are now approaching 100 Gbps, using both copper and fiber media.  Space doesn’t permit the details, but are located at the IEEE website at www.ieee.org.

Knowing how Ethernet operates is critical to success as a network professional, as this understanding comes in remarkably handy for operations and troubleshooting purposes.  The first building block is knowing that all devices are connected to the network media (copper or fiber) and having equal access to available resources.  This includes both transmitting and receiving of data, but naturally there is an arbitration mechanism in the event of multiple devices trying to do that at the same time.  This is referred to as Carrier Sense Multiple Access with Collision Detect (CSMA/CD), and is best illustrated with the following link: http://www.metacafe.com/watch/1161647/intersection_full_of_accidents/.

The process is simple: A transmitting station listens to make sure the “road” is clear,and if there is no traffic, they send the data.  If, however, another station is also trying to transmit, the result is  lot like a car accident, as the frames collide and become useless.  In an accident, police and/or DOT arrive on the scene, clear the blockage, and get things moving again.  In Ethernet, the affected stations send out  jamming signaling, causing all traffic to stop, and then wait to transmit again, using a random timer.  Next time we will look at cabling.

– Joe

CCNA Networking Fundamentals: Part 3/Last

Posted in Cisco Certification with tags , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , on April 8, 2011 by jjrinehart
 One of the other important exam elements–as well as networking foundations–is a discussion regarding the OSI (open systems interconnection) reference model.  Most technical documentation makes reference to this model, so knowing it backwards, forwards and slantwards is critical to showing competence.  Back in the mists of time, every equipment manufacturer built their own systems, components, and cabling; this is referred to as proprietary, meaning non-standard.  In simple terms, it means that not only do you have to buy anything and everything from that vendor, but also only that vendor’s equipment will communicate with that same equipment.  This was fairly typical in the old mainframe days, as well as when the PC first came out.  For instance, back in the 1990’s I bought a Tandy 1000 TL computer, and when I wanted to add another floppy drive, I was stuck because of the way Tandy made the cables.  Packard Bell computers were the same way…cheap to buy, expensive to upgrade because of the proprietary components.

Reference models began to appear at this point, with OSI, TCP/IP, and Novell being among the most prominent.  In Wendell Odom’s CCNA Certification Library, he illustrated this as follows:

Reference Models

Novell, like Sony Betamax, eventually became obsolete over time, much like ARCnet, Token Ring, and IPX.  The OSI model, on the other hand, has been burned into the proverbial geek psyche and the common reference point for network technologies.  Know the layers, in order, in your sleep, as well as what each layer does.  For example, in discussing switches, you will hear references to Layer 2 (data link) switches, as opposed to Layer 3 (network) switches, describing how they process network traffic (more on this later).  Here is a common breakdown of each layer in the stack, along with the typical data units involved, and the protocols in use:

OSI Model In Detail

This sets the stage for the rest of the course of study for the CCNA exam, as my own lessons (as well as this blog) will start at Layer 1 (physical layer) and move up the stack in a logical fashion.  Next time (Monday), we will begin wading through the world of Ethernet and Local Area Networks!

CCNA Networking Fundamentals: Part 3

Posted in Cisco Certification with tags , , , , , , , , , , , , , , , , , , , , , on April 7, 2011 by jjrinehart

Today we will look a little more closely at addressing, specifically network addressing (there are more categories we will discuss later on).  When you look at an Internet Protocol (IP) address, it can seem like a totally foreign language, mostly because it is not readily familiar.  One word to get used to with IP addresses is octet, which means “eight of something”, or “group of eight.”  The reason for this is that the smallest unit of the addressing, based on binary, is a bit (remember, which is either a 1 or 0), and there are eight bits in a byte.  Each decimal number is  byte or octet, hence the meaning of the word in this particular context.  Here is an example of an actual IP Address: 24.6.12.41. 

Each part of the address has significance, and refers to some part of a larger group of addresses belonging to some organization, which will also be part of a later discussion.  The hierarchy of numbers is similar to the way telephone numbers operate (it should be noted that with local number portability and cellular phones that some of these lines are blurred nowadays).  A random phone number illustrating this is as follows: 425-519-6476.  One of the somewhat invisible aspects of this set of numbers is the country code, which does not need to be used within the borders of a given nation, the US code is 1.  The whole number breakdown, with associated geography, map out this way:

1: United States

425: Washington State, Seattle Eastside Suburbs

519: City of Bellevue

6476: Unique/Individual Subscriber Number

When a person in another state pucks up the phone and dials the complete string of numbers above, the equipment knows how to send the call based on each of these numbers, moving from the most general (country) to the most specific (subscriber).  We do this often enough that at times we even know where someone is dialing to/from based on the area code.

Now let’s apply this to IP addressing, which to some extent is a little simplified, but still largely makes sense.  Here is the same type of breakdown of the phone number used previously, using the IP address 24.6.12.41.

24: Comcast Cable Communications (you can look this up at http://whois.arin.net/rest/net/NET-24-0-0-0-2/pft)

6:  Comcast San Francisco Bay Area

12.41:  Unique Customer Address

IP address usually follows this type of hierarchy, which is how Internet routers know where to send traffic.  Getting familiar with this type of concept will save you  lot of grief later, and help you in applying these concepts to the CCNA exam as well as in practical settings.

Next time we will review another foundational topic, the Open Systems Interconnection (OSI) Model

– Joe

CCNA Networking Fundamentals: Part 2

Posted in Cisco Certification with tags , , , , , , , , , , , , , , , , , , , on April 5, 2011 by jjrinehart

As we discussed last time, networking technologies are simply another form of communication, namely, getting information from one entity to another.  Let’s take the basic communication model one step further.  The Internet (the ultimate network) is often called the “information superhighway” so using that analogy can also be helpful.  Whenever we travel, there are fundamental building blocks of how that process happens.  We have a starting point (usually a physical address) and of course an end point in mind (also a physical address).  Once we know where we are going, we get into a vehicle and take a ramp on to the highway, and eventually an off ramp to the street we want to get to.  After we arrive, we park, get out of the car, and enter into that location.  The following diagram illustrates this process, one that we literally use many times every day.  Simple enough, right?

Information Superhighway

Getting from One Place to Another on the Highway

Now let’s map this analogy back to the world of CCNA/computer networking.  Information/data needs to get from one location to another location some distance away. Let’s alter the previous diagram a little bit to show this process more accurately:

Highway Analogy Mapped to Networking

The computer encodes (just as in conversations, discussed earlier) the data, and records its own location (source address) in the encoding process, as well as the location it wants to send the information to (destination address).  Pictured here are two randomly chosen addresses to illustrate the point.  At this point the information is inserted into a packet (a vehicle for transportation), where it enters the highway through an access link (like an on-ramp), an travels the road (network) until it nears the destination,and exits as close to that location as possible (another access-link, like an exit ramp).  Finally, it arrives at the intended, and the packet is stripped off (like leaving a vehicle), and the data is delivered.  Just as on the actual highway, the vehicle had to obey the necessary traffic laws (protocols) to successfully navigate the traffic medium.

That’s basically networking in a nutshell, I came up with this way of explaining it years ago, to help take the mystery out of the process.  Everything else we will discuss can easily map back to this analogy.  Next time, we will take a look at network addressing before diving into some deeper topics.