By: Yigal AmramSpine-and-leaf data center architecture has been a major benefit to network designers trying to keep up with ever-increasing business demands. You’re tasked with supporting data-heavy applications, bandwidth-hungry communications systems, and more connected endpoints—all with fewer resources to scale the data center. And while spine-and-leaf designs have helped us get by, we’re quickly reaching the limits of their scalability.
If you want to balance scalability with the performance necessary for modern business demands, you can’t keep pushing the limits of standard spine-and-leaf designs. With single-mode fiber cables and port fan-out technology, you can design next-generation networks with maximum visibility.
The Network Design Flaw with Single Mode Fiber
Fiber optic technology has advanced to the point where 40 Gigabit Ethernet is readily available for your network designs. In theory, single mode 40G fiber optic cabling is the perfect solution to your performance issues. It supports high bandwidth levels, minimizes signal loss, and essentially has no distance restrictions. And as technology continues to evolve, single mode fiber is becoming increasingly cost effective for short-distance applications.
The only problem is that you most likely aren’t building your network from scratch. Upgrading an existing spine-and-leaf design with single mode 40G fiber cabling has its limitations. Because most common top of rack (ToR) leaf switches come with four 40G uplink ports, you’re locked into four spine switches. And with each spine containing four line cards supporting 36 40G ports each, the total number of available leaf switch ports is 144. Then, each ToR leaf switch has 40 ports to connect to servers, giving you a maximum of 5,760 servers to connect to a 40G mesh network.
These numbers may have worked for your network design in the past. But what happens when you need to scale beyond this maximum limit? This is the flaw with single mode fiber optic cabling. For all its performance benefits, single mode 40G fiber won’t overcome the scalability challenges for the traditional spine-and-leaf design.
That’s where fan-out technology comes into play.
Scaling 40G Single Mode Fiber with Port Fan-Out
Port fan-out technology can help you take advantage of the performance benefits of 40G single mode fiber while making your data center more scalable.
With fan-out single mode fiber cables, you can turn single 40G uplinks into four separate 10G links. This way, instead of four 40G uplink ports, you have 16 separate ports, increasing the number of spine switches that can fit in your 40G network mesh. Instead of the 5,760 server limit with standard single mode fiber, you can multiply scalability by four and support up to 23,040 servers.
It’s the same hardware, bandwidth, latency, and oversubscription you get from a traditional spine-and-leaf design without the scalability issues. Additionally, fan-out technology makes it easier to maintain interoperability between existing 10G appliances and the upgraded 40G single mode fiber optic cabling you’re installing. Instead of worrying about media conversions and an array of different connectivity solutions, you can simplify your network design while still addressing the growing demands of your business.
However, a new problem emerges when you start increasing port density with fan-out technology—maintaining visibility for your security and monitoring tools.
3 Ways to Maximize Network Visibility for Port Fan-Out and Single Mode Fiber
Next generation networks must support an increasingly diverse set of applications and infrastructure services. But the result of that demand is an ever-growing need to deploy visibility and monitoring solutions that streamline troubleshooting for physical and virtual infrastructure.
Deploying 40G single mode fiber optic cables puts you on a path to creating this kind of next-gen network. However, it must be accompanied by a robust visibility adaptation layer if you want to get the most out of your security and monitoring tools. With the right combination of network taps, bypass switches, and network packet brokers, you can build pervasive visibility into your data center fabric.
Packet brokers, in particular, play an essential role in your ability to leverage port fan-out technology with 40G single mode fiber. There are three key use cases where a packet broker should support port fan-out, including:
- Scalability: A modular network packet broker can use port fan-out for 40G single mode fiber across bypass and packet broker modules to optimize for rack space and enhance the efficiency of your monitoring, analysis, and security tools. This is the use case discussed above, where you can increase density by turning a 40G port into four 10G ports.
- Bypass: Port fan-out support in a network packet broker enables 40G appliance ports in a bypass module to break out and operate as separate 10G ports. This creates distinct network segments that can connect to individual 10G appliances. You get to extract more value from your existing 10G appliances while ensuring that in-line devices don’t become points of failure in a scaled data center network.
- Packet Brokering: Port fan-out helps you support monitoring, analysis, and security tools with lower throughput capacity. Together with load-balancing capabilities in a packet broker module, you can ensure no port is oversubscribed and that packet visibility remains at 100%.
Scalability, performance, and visibility. These are the key pillars of a next-gen network. And while adopting 40G single mode fiber may seem like the key to achieving your goals, it can’t address all three pillars on its own.
Creating a network visibility layer that leverages port fan-out technology for your single mode fiber optic cabling will give you the foundation for a next-gen network. But that’s often easier said than done.
If you want to learn more about maximizing the ROI of your data center design with the help of port fan-out technology, contact us today to talk to a Niagara Networks visibility expert.
