If your organization is consolidating or redesigning data centers, even building a new one, or modernizing your campus network, you may be thinking about new ways to optimize the switching infrastructure. Certainly, high availability, performance, operational simplicity, and cost remain key considerations. However, scalability and power consumption may be top of mind as the challenges you most want to solve.
Downside of a three-layer configuration
In the data center or campus network, the access layer provides the physical connections to servers, storage, security, and other IP devices. In a typical configuration, aggregation switches interconnect these access switches, while a core layer provides connectivity between the aggregation layer and the gateway routers that link to the Internet and/or the WAN that interconnects all your sites.
While this three-layer breakout allowed new devices and switches to be added without requiring a major overhaul to the existing network, in many data centers, more than 50 percent of available switching ports are now used to connect to other switches.
The compounding complexity of scaling across three layers not only adds wiring but also can increase the risk of failure and the effort to manage the infrastructure. Also, the power and space needed to run all those switches can further strain energy and financial budgets.
For many organizations, the challenges are only growing with not only more and more traffic but new requirements driven by cloud-based services and virtualization. For example, virtualization adds the need to move much of the traffic across servers in the data center, requiring far more flexibility than when the network simply had to move the traffic in/out of dedicated application server farms.
If your applications and traffic needs are changing, you may be looking at design alternatives for your data center and campus networks. New switching solutions with higher port densities, faster interface speeds, and more flexible ways to expand ports offer new options for optimizing your network for growth.
Advantage of collapsing to a two-layer network
A design option that you may be considering to optimize your data center or campus network is reducing the number of switching layers. The availability of high-density switches with many high-speed 10 Gigabit Ethernet interfaces allow you to consider collapsing the aggregation and core layers of your network(s). In this design, your access switches can directly connect to the core over wire-speed 10 Gigabit Ethernet links.
By reducing the number of switches, the design can save not only capital cost but also reduce power, cooling, and space requirements. Additionally, collapsing layers removes potential points of failure and simplifies network operations — including OS upgrades; moves, adds, changes; and troubleshooting.
If you are upgrading your switching infrastructure, you may begin your migration to Junos by deploying new switches that can eliminate the aggregation layer of your network.
Virtual Chassis technology for collapsing layers
The Juniper solution provides a way for a group of interconnected switches to operate as a single, logical device with a single IP address. The key enabling capability is known as Virtual Chassis technology. When deployed in a Virtual Chassis configuration, the switches are also monitored and managed in Junos OS as one device, enabling organizations to separate physical topology from logical groupings of endpoints and allowing more efficient resource utilization.
Virtual Chassis technology not only reduces the number of individual devices to manage, it also allows network tiers to be consolidated, further simplifying the network. A single Virtual Chassis configuration can also span multiple nodes, reducing the need for uplinks and limiting the need for larger, more expensive nodes on superior tiers.
dummies
Source:http://www.dummies.com/how-to/content/collapse-switching-layers-with-junos-to-optimize-s.html
No comments:
Post a Comment