Changes in data center design will affect network switches

Comments · 63 Views

The data center level switch incorporates a wider range of high availability and fault tolerant systems in both hardware and software, providing better uptime for mission critical applications. They provide higher deployment flexibility and compatibility with rack top and row end configura

Core switches

The core switch is the easiest to understand. The core of a core network switch is speed. If properly designed, the only task that the core switch should perform is the routing of Layer 3 (network layer) and the switching of Layer 2 (data link layer that moves data on the physical link of the network).

Core switches are high throughput, high-performance packet and frame mobile devices. Packets and frames simply move from one core switch to another and ultimately down to the next layer of the switch - the distribution layer.

Power distribution switches

The distribution layer addresses a new set of unique exchange requirements and is the backbone of any enterprise network.

Firstly, the distribution switch is used to connect the core layer and access layer in the network. If data needs to be moved from one distribution block to another, the switch pushes the data up to the core switch, which knows the best path to reach the target distribution layer switch.

Secondly, the distribution switch also interconnects all network access layer switches. Due to the large number of interconnections in the network, the port density of allocated switches is higher than that of core switches, and the interconnections between core switches and other switches are much less.

Finally, and most importantly, allocate switches to execute all forms of network policies. Access lists are configured and implemented in the distribution layer to allow or deny traffic from one network to another. A quality of service strategy was also discovered here to prioritize data packets and place them in predefined queues for optimal transmission of time sensitive information. In addition to port density, distributed layer switches must also have sufficient CPU speed and memory to perform all tasks at or near line speed.

Access switches

At the bottom of the classic three-layer switch design is the access layer.

The access layer switch is the only switch that directly interacts with end user devices. Due to the fact that access switches connect most devices to the network, the access layer typically has the highest port density among all switch types.

Although the number of ports is high, access switches typically provide the lowest throughput per port among all switches. For example, most modern access switches provide 10/100/1000 Mbps copper Ethernet connections for terminal devices. In contrast, core and distribution layer switches typically use fiber optic ports ranging from 10 Gbps to 100 Gbps.

Therefore, in terms of CPU and raw throughput, the access switch is at the low end. But these switches provide many functions specifically for terminal devices that are not needed in the upper layer. For example, access switches usually support power over Ethernet, which can power many endpoint devices, including wireless access points and security cameras.

In addition, from a security perspective, access switches can better interact with endpoints. Port security, 802.1X authentication, and other security mechanisms are all directly established on the access switch software.


Changes in data center design will affect switches

The three-layer design connects devices and layers across enterprise infrastructure, with high scalability. Until a few years ago, only access switches were used to connect servers to other parts of the network, using the same 1 Gigabit Ethernet (GbE) ports as desktop computers or networked printers. However, over time, changes in data center server architecture - taking the continuous development of storage area networks (SANs) and virtualization as the pace - have spawned a new high-performance switch, which we refer to as a data center level switch.

These switches provide the physical port capacity and port throughput required to handle north-south and east-west traffic. They allow connections to be made using standard LAN Ethernet protocols and SAN protocols, such as Fibre Channel over Ethernet and traditional Fibre Channel. The data center level switch incorporates a wider range of high availability and fault tolerant systems in both hardware and software, providing better uptime for mission critical applications. They provide higher deployment flexibility and compatibility with rack top and row end configurations. Finally, all components of distributed data center switches can be managed through a single management interface for ease of use.