your123 - stock.adobe.com
These days, a switch is a switch, right? Well, not necessarily. While all network switches move data from point A to point B, many enterprise-grade switches have significant hardware and software differences that factor into any IT infrastructure deployment. Let's look at the different types of network switches available on the market today.
1. Unmanaged, smart and managed switches
Understanding the level of management and configurability of a network switch is one of the most important buying decisions a network architect must make. For small branch locations or work-from-home offices, an unmanaged switch may fit the bill. These switches are essentially plug-and-play units that enable multiple devices to communicate across a single broadcast domain. Because of their limited capabilities, unmanaged switches are considerably less expensive compared to smart and managed switch alternatives.
When comparing the differences between smart and managed switches, things start to get a bit murky. Both devices are technically manageable devices. However, in most cases, network equipment vendors that sell smart switches tend to strip out many of the more advanced features and only include the basics, such as virtual LAN (VLAN) creation, basic quality of service settings, port aggregation and a few Spanning Tree Protocol options. Smart switches are commonly configured through a web GUI, as opposed to a command-line interface (CLI).
Managed switches, on the other hand, are at the top of the switch food chain. These switches offer hundreds to thousands of configuration options -- many of which are highly useful for medium- to large-sized corporate LANs. Additionally, management of these devices can include a GUI -- but, more often, they are managed via CLI for speed and ease of use by trained network professionals.
2. Layer 2 and Layer 3 switches
Keeping the focus on managed switches, these can be further segmented into two distinct feature types. They are typically referred to as Layer 2 and Layer 3 switches based on where they operate on the OSI model. Layer 2 switches are also referred to as multiport bridge switches, while Layer 3 switches are sometimes called multilayer switches.
Layer 2 switches can intelligently move data frames from one port to another on the same VLAN. However, data that needs to move between VLANs -- also known as inter-VLAN routing -- needs a device that can route IP packets. When using Layer 2 switches, this step is often done with an external router using a one-armed architecture.
For large networks with multiple VLANs and a lot of routing between them, it's often easier and more efficient to combine the capabilities of a Layer 2 switch and a router into a single hardware and software device. This is precisely what a Layer 3 switch does.
Instead of relying on an external device to route traffic between VLANs, a Layer 3 switch can be configured to do this across its own internal switching backplane. Thus, for LANs that require a routing component, a Layer 3 switch reduces the network equipment footprint and increases performance compared to one-armed designs that rely on an external routing component.
3. Power over Ethernet switches
Power over Ethernet (PoE) is the ability to send low-voltage electricity across the same twisted-pair copper cabling that is used to transmit and receive data. This feature is used to power wireless access points (APs), IP phones and numerous IoT devices.
If PoE is not required in any capacity, then non-PoE switches are a cheaper option. However, for those that need PoE, a few additional steps are needed to ensure PoE endpoints receive sufficient power.
PoE standards, as dictated by IEEE, specify the maximum wattage that can be transmitted across copper cabling. Depending on the endpoint, more or less power may be required. For example, a typical IP phone can be powered via PoE using a PoE port that can transmit up to 15.4 watts (W) of power. On the other hand, modern Wi-Fi 6 and Wi-Fi 6E APs may require substantially more power to operate. Thus, a PoE switch that is only capable of delivering IEEE 802.3af standards will not suffice.
The following is a list of PoE standards and the maximum wattage that can be delivered across the recommended category of twisted-pair cabling:
|PoE standard||IEEE 802.3af||IEEE 802.3at||IEEE 802.3bt (Type 3)||IEEE 802.3bt (Type 4)|
|Maximum wattage||15.4 W||30 W||60 W||90 W|
|Recommended cabling||Cat3 and Cat5||Cat5 or better||Cat5 or better||Cat5 or better|
4. Fixed, modular and stackable switches
From a physical standpoint, network switches come in three different hardware configuration types:
- Fixed switches. With fixed switches, ports, interfaces, power supplies and cooling fans are set and cannot be changed, added or altered. Additionally, fixed switches cannot be stacked onto other switches to create a single logical switch from which to manage.
- Stackable switches. Stackable switches are fixed switches that include a backplane cable interface to connect multiple switches together to create a single logical switch made from two or more physical switches. Doing so can increase the speed of switch-to-switch data transport, as well as simplify the management of the stack as several physical switches are managed as if they were one single switch. Some stackable switches can also share power between each stack. Thus, if a switch in the stack suffered a failed power supply, it can continue operating by taking unused power capacity from other switches in the stack.
- Modular switches. Modular or chassis-based switches offer the ability to insert switch cards into a large, fixed-form factor chassis that can support two or more cards. This type of switch offers the most flexibility and upgradability as switch interfaces can be swapped out as needed. Additionally, if a card failed on a modular switch, a field technician can hot swap the failed card out without bringing down the entire switch. Lastly, it is common for modular switches to be able to swap out power supplies and cooling fans when upgrades are required or failures occur.
Dig Deeper on Network Infrastructure
Related Q&A from Andrew Froehlich
Companies can take some simple steps to protect their Zoom meetings, but they should also consider ways to better educate their employees on how to ... Continue Reading
An important piece of remote and hybrid work is keeping UC environments secure. Learn how a zero-trust policy keeps user and app communications ... Continue Reading
The difference between network redundancy and resiliency is redundancy duplicates network devices while resiliency is the self-recovery of system ... Continue Reading