Network Switch Selection – How to Select a Network Switch

Overview

The network switch is the most common network device implemented with company infrastructure and as such the selection of any new switches or upgrading is a key part of most network design projects. The Cisco network switch components include Switch Chassis, Supervisor Engine, Switching Modules, IOS/CatOS software and Power Supplies. The decision to buy new switches or upgrade equipment will be decided after considering the network assessment and design features specified. Wireless designs, as an example, will have network switches interfacing with access points. That will have an affect on the switch such as increased utilization, assigned switch ports, access control lists, Trunking, Spanning Tree Protocol and increased wattage draw from Power over Ethernet (PoE).

Switch Chassis Features

The Switch Chassis features include – chassis dimensions, number of slots, processor slot assignments, switching fabric, engines types supported, power supplies, rack units needed.

Cisco Supervisor Engine (SE) Features

Cisco switches are implemented with an Engine (Switch Processor) for processing packets on a network segment. Routing is accomplished with an on-board Multi Layer Switch Feature Card (MSFC) or Route Processor running IOS code. The switch Engine running IOS code on the MSFC and the switch processor is in native mode, while those running CatOS on the processor are in hybrid mode. Some engines won’t support native and hybrid mode. The engine with no MSFC supports what is called CatOS mode. Select the engine that matches your design specifications. The MSFC module is integrated with the Engine or upgradeable. You must implement a PFC module with any MSFC. Some Engines have no MSFC module – the routing is integrated with the hardware and as such support native mode only.

The Cisco Supervisor Engine features include – supported chassis, uplink speed, processor memory, native IOS, CatOS, PFC, MSFC, slot assignment, failover.

These are some of the popular Cisco engines and their switching features.

720 – Cisco 6500 switches, 400 mpps, MSFC3, IOS, CatOS

32 – Cisco 6500 switches, 15 mpps, MSFC2A, IOS, CatOS

V – Cisco 4500 switches, 72 mpps, Integrated Routing, IOS

IV – Cisco 4500 switches, 48 mpps, Integrated Routing, IOS

Switching Module Features

The Switching Module features include – supported switch chassis, interface speed, number of ports, media, cabling, connectors, throughput (mpps), supervisor engines supported, protocol features, power over ethernet (Cisco prestandard or 802.3af).

– Media: Copper, Fiber

– Cabling: UTP Cat 5, CAT 5e, CAT 6, STP, MMF, SMF

– Connectors: RJ45, RJ21, SC, LC

– Transceivers: GBIC, SFP

Power Supply Features

The Power Supply features include – supported chassis, wattage ratings, failover, input/output amps, power cord type, IOS, CatOS.

IOS/CatOS Software

Cisco network switches can be deployed with IOS, IOS and CatOS or exclusive CatOS software. Design features will determine what mode and IOS or CatOS version is selected. The software running on the Route Processor must be IOS while the Engine Switch Processor will run IOS (native mode) or CatOS (hybrid mode). Some Cisco equipment such as the 4507R deploy the Supervisor Engine IV with no MSFC onboard. The Route Processor is integrated with the engine. With that design, the Engine IV doesn’t support CatOS.

Native IOS – deployed at the network edge where most routing occurs and some switching is needed

Hybrid – deployed at the network core where there is both routing and high speed switching

CatOS – deployed at the network access layer where there is switching and no routing

Switch Selection Process:

The following describes the 5 components of any network switch selection process:

1. Consider the network assessment and design features specified

2. Select switches that include all the design features

3. Select switches with proper scalability

4. Balance cost and equipment features while meeting budget guidelines

5. Select IOS and/or CatOS software version

The Network Assessment and design specifications should be considered before selecting any network switches. The network assessment examines the design, configuration and equipment that is implemented at the office where the selected devices will be deployed. The design specifications will determine performance, availability and scalability features needed. Selecting the IOS and/or CatOS version occur after deciding on the feature set. Companies will specify a budget and that is a key consideration with any equipment selection. It isn’t cost effective to select a Cisco 6509 switch for an office with 50 employees. It is important that you select equipment that meet the design specifications, has the scalability features needed while meeting budget guidelines.

Some typical switch features to consider:

1) Are there enough Chassis slots?

2) What Supervisor Engines are supported?

3) Does the Engine support failover?

4) Is Multi Layer Switching available?

5) What Switching Modules are available?

6) What Uplinks are available?

7) What Power Supply wattage is available?

8) How many Rack Units are needed?

Switch Selection Example:

The Network Assessment discovered the following at the company office.

· The Distribution Office has 300 employees

· Fast Ethernet (100BaseT) is implemented at the Desktop

· 6509 Cisco Switches with Gigabit Ethernet Trunking

· 3800 Cisco Router with dual T1 Circuits

· Power over Ethernet is implemented

· Multiple VLANS defined

· Local Unix and Windows Servers

· Some bandwidth intensive applications

· IP Telephony is implemented at all offices

· Wiring closets are 500 feet apart

· Several Rack Units are available in the Rack Cabinet

The design specifies that an additional 180 people will be employed soon. The company will have those employees working from a third floor where the nearest wiring closet is 500 feet from the Cisco 6509. The company will implement some Wireless, IP Telephony and define VLANS with each specific company department.

The following is a list of specific switching features needed:

1. 4 Chassis slots with Switching Modules of 48 Port – 10/100BaseT

2. Gigabit Ethernet Trunking between wiring closets

3. Supervisor Engines with failover

4. Multi Layer Switching

5. Power over Ethernet support

6. Dual Power Supplies with at least 2800W for IP Phones

7. Quality of Service for IP Telephony

8. Performance switching for converged telephony network

Selected Switch: Cisco 4507R

The Cisco 4507R switch has 7 slots and is a good selection with the additional 180 employees. The device will have 4 – 48 port modules with a slot available for any additional employees. The dual Engines IV will be assigned 2 slots with failover, Multi Layer Switching between VLANS and Gigabit Ethernet uplinks connecting the 6509 devices. Each of the switching modules are PoE capable with the new 802.3af standard. Dual power supplies provide enough wattage for implementing hundreds of Cisco IP Phones and Wireless Access Points. The engine performance is 75 mpps with wire speed switching. The Cisco 4507R is more cost effective than the nearest Cisco 6509 device. Several Engine models are available with additional performance features.

– 7 slot chassis with 2 Supervisor Engines and 4 Switching Modules with 48 Port – 10/100BaseT

– Engine IV with integrated Multi Layer Switching, failover, dual Gigabit uplinks

– Power over Ethernet (PoE) support with 802.3af standard

– Dual Power Supplies with 2800W or 4200W for Telephony, Wireless, Power over Ethernet

– Quality of Service features for voice traffic

– Fast performance with 75 mpps wire speed switching for converged networking

The 3750 series Cisco switch wasn’t as expensive however there were not enough slots, stacking technology is expensive and switches at 38.7 mpps compared with the 4507R device at 75 mpps. The company would have to buy 5 separate switches with 48 ports for 180 employees. The Cisco 2950 switch doesn’t have power supply failover and scalability. The 6509 switch was much more expensive, had 2 additional slots, more performance than was needed and the switching modules were expensive. Implementation is somewhat difficult with the 6500 Cisco devices.

Network Switch – The Basics

The network switch plays an integral role in enterprise and home networking, yet many people confuse what the purpose of the equipment is, and how it differs from a router. I decided to write this blog post to explain the basics of the switch – from different types, to vendors for purchasing them.

A brief overview of network switches

A network switch is a type of computer networking hardware that bridges network segments. It is sometimes referred to as a packet switch or simply a switch. The switch plays an important component in most local area networks (LAN), including mid-to-large enterprise networks which utilize several linked managed switches.

A switch is far less sophisticated than a router. Although routers and switches look fairly similar in appearance, routers differ substantially in their internal components.

Types of network switches

Unmanaged Switches: This is typically the least expensive type of switch, most often found in homes or small offices. They are very simple, employing plug and play technology, lacking any specific configuration options

Managed Switches: Managed Switches provide optional configuration options and allow for a great variety of functionality. There are several ways to operate these switches, from utilizing a remote tool like Simple Network Management Protocol (SNMP), to accessing the switch via a command line interface like Telnet.

  • Smart Switches: Smart switches differ from fully managed switches in that they only allow a specific set of modifications and functionality. Because users can only configure basics settings, they are often cheaper than the fully managed breed. Some basic functions often found on a smart switch are turning some particular port range on or off, link speed and duplex settings and priority settings for ports
  • Enterprise Managed Switches: Enterprise switches are the more configurable and expensive version of managed switches. They are most often found in enterprise networks among several other switches. They are more efficient for large business where accessing a central administration module can save time and money. Some advanced functions for enterprise switches are VLAN settings, link aggregation and port mirroring.

Buying switches

There are several brand name switch manufacturers that provide competing and differentiated products, including Cisco, 3Com, and Alcatel. While switches can be purchased out of the box from online retailers, one way to save money is to find a used switch from an online reseller. A business purchaser can often save thousands of dollars purchasing used cisco or other brand name network hardware.

If you do decide to go the route of an online reseller, be sure to check for several qualifying factors to make sure they are a good fit. One factor is a good warranty, as it is always a risk to buy used equipment. Another is significant discounts (at least 50%) off of retail pricing. The third factor I recommend seeking in an online network hardware vendor is good customer support. The ability to speak to a human being for help with your purchase is underrated.

I hope this ‘basic switch support’ post helps out those that are confused or looking for a way to purchase a switch.

The Truth About Current Switching and Measurement Systems

Unless signal differences are taken into account, they can degrade signal integrity and affect total test system performance

The differences among the types of signals that a test system’s switching hardware must handle are not always well understood. But if these differences are not taken into account in switch system design, they can degrade signal integrity and affect overall test system performance.

When designing a measurement system, selection of the switch is as critical as the selection of system instrumentation or the design of the test interface. The intended application must be thoroughly considered, and the switch selected must meet the requirements of the application. Careful attention to detail and to the basic principles of measurement can help ensure greater system accuracy and performance.

Voltage vs. current switching
Voltage sources can usually provide a compliance current up to the programmed voltage. As a result, the typical default condition of a voltage switch is open (in other words, drawing very little current or having high impedance).

Current switching, however, usually requires the default configuration to be a complete circuit. This means the current needs a complete path until switched. Typically, the switching component (that is, the relay) is a normally closed relay, or the HI and LO terminals are shorted in the default condition. A variety of switching topologies are suitable for use in current switching applications: scanner, multiplex, and matrix switching. The scan configuration or scanner is the simplest arrangement of relays in a switch system. It can be thought of as a multiple position selector switch.

Like the scan configuration, multiplex switching can be used to connect one instrument to multiple devices (1:N) or to connect multiple instruments to a single device (N:1). However, the multiplex configuration is much more flexible than the scan configuration. Unlike the scan configuration, multiplex switching enables making multiple simultaneous connections and also permits either sequential or nonsequential switch closures.

The matrix switch configuration is the most versatile because it can connect multiple inputs to multiple outputs. A matrix is useful when connections must be made between several signal sources and a multi-pin device, such as an integrated circuit or a resistor network.

Typical current concerns
Most current measurement applications demand that all current paths be continuous, even when a particular current signal is not connected to the ammeter. To accomplish this, switch cards designed for current switching often use SPDT or Form C relays.

Note that the current will be interrupted briefly when the Form C relay is actuated. This could cause problems when used with high-speed logic or other circuits sensitive to a momentary break in the current flow. Such a problem can be overcome by using a switch card such as those used with the Series 3700 switch system/multimeter with a pair of Form A isolated switches to provide a make-before-break connection.

High-current considerations
When designing a switching circuit for high current (> 1 A), pay particular attention to the maximum current, maximum voltage, and VA specifications of the switch cards and relays. Also, it is important to choose a switch card or relay with low contact resistance to avoid excessive heating, which can cause contacts to weld together and thereby lead to contact failure. Contact heating is caused by I2R power dissipation.

High-current switching can be used for either switching a power supply to multiple loads or for switching an ammeter to multiple sources. When a power supply is switched to multiple loads using a multiplexer scanner card, the power supply will output 1 A to each of four loads. This does not present a problem when only one channel is closed at a time, but when all four channels are closed, the power supply will output 4 A through the common path.

Unfortunately, even though the maximum current of a particular channel is specified at 1 A, the common path on the switch card may not be able to tolerate 4 A. This is not usually specified for a switch card, but the limitation is usually a function of the trace width and connector ratings. One way to avoid this problem is to use a switch card with independent (isolated) relays and to make connections with wires rated to carry the total current.

Low-current considerations
When switching currents of 1 micro-A or less, special techniques must be used to minimize sources of interference such as offset currents, leakage currents, electrostatic interference, triboelectric currents, and electromechanical currents. The interference might come from the switch card itself, the connecting cables, or the test fixturing.

Offset currents are spurious currents generated by a switching card even though no signals are applied. They are usually caused by galvanic sources on the switch card. Offset current is especially significant when measuring low currents if the magnitude of the offset is comparable to that of the current being measured.

Leakage current is an error current that flows through insulators when a voltage is applied. It can be found on the switch card, in cabling, and in test fixtures. Even high-resistance paths between low-current conductors and nearby voltage sources can generate significant leakage currents.

To reduce these effects, always use a switch card with high channel isolation and use the guard capability of the measurement instrument. Another method to help reduce leakage current is to keep the switch card clean. Dirt, body oils, and the like will create a lower-resistance path and allow leakage currents to flow.

To reduce leakage currents in the text fixturing, always use good quality insulators such as Teflon and polyethylene; avoid materials such as nylon and phenolics, which can absorb moisture, affecting their insulating performance.

Shielding is required because high-impedance circuitry is susceptible to picking up spurious radiated noise. Relay contacts should be shielded from the coil to minimize induced noise from the relay power supply. The device under test and interconnect cabling should also be shielded to prevent noise pickup. All shields should be connected to circuit LO.

Triboelectric currents are generated by charges created by friction between a conductor and an insulator, such as between the conductor and the insulation of a coaxial cable. The friction can be reduced by using special low-noise cables with conductive coating (such as graphite) and by securing the interconnect cabling to minimize movement.

Electrochemical currents are generated by galvanic battery action that results from contamination and humidity. Cleansing joints and surfaces thoroughly to remove electrolytic residues (which include PC etchants, body salts, and processing chemicals) will minimize these parasitic battery effects.

Settling time
When a relay opens or closes, a charge transfer on the order of picocoulombs occurs, which causes a current pulse in the circuit. This charge transfer is due to the mechanical release or closure of the contacts, the contact-to-coil capacitance, and the stray capacitance between signal and relay drive lines. After a relay is closed, it is important to allow sufficient settling time before taking a measurement. This can be as long as several seconds, depending on the relay.

If a step voltage is applied to the circuit, a transient current is generated. This current will gradually decay to a steady-state value. The time needed to reach the steady value (or settling time) can be used to determine the proper measurement-delay time.

Cold, hot, and safe
The term “cold switching” indicates that a switch is activated with no signal applied. Therefore, no current will flow when the switch is closed and no current will be interrupted when the switch is opened. When hot switching, voltage is present and current will flow the instant the contacts close. When the switch is opened, this current will be interrupted and can cause arcing.

Cold switching lets power be applied to the device under test in a controlled manner. Its primary advantage is longer switch life than with hot switching. Cold switching also eliminates arcing at the relay contacts and any RFI it might cause.

Hot switching might be necessary if close control must be exercised in the period between the application of power and the making of the measurement. For example, hot switching is typically used where digital logic is involved, because devices might change state if power is interrupted even for a moment. With relatively large relays, hot switching should also be performed every so often to ensure good contact closure. The connection might not be reliable without the “wetting” action produced by current flow through the contacts.

Many electrical test systems can produce hazardous levels of power. These high power levels make operator protection a priority. Some protection methods include:

– Designing test fixtures that prevent operator contact with hazardous circuits
– Double-insulating all electrical connections that an operator could touch
– Using high-reliability fail-safe interlock switches that disconnect power sources when a test fixture is opened
– Providing proper training to all users so that they understand potential hazards and know how to protect themselves from injury