The reign of proprietary networking in optical and data center interconnect falls

In furthering our mission to bring S.O.U.L. to networking through Simple, Open, Untethered, Linux-based networking solutions, and just on the heels of our Cumulus NetQ 1.3 announcement around simplifying container networking and operations, today we continue the mission to advance web-scale networking in the digital age with our release of Cumulus Linux 3.6. Our focus has been to help organizations move towards a modern world of simplification, flexibility and scale — where complex applications reside on standardized infrastructure that is automated, repeatable and scalable. We see a world of agility built upon cloud principles; of converged administrative teams where sysadmins can manage the network and network admins can manage systems.

What’s new in Cumulus Linux 3.6

In this release of Cumulus Linux 3.6, we are not only driving network efficiency and simplicity, but also expanding our solution set to include data center interconnect (DCI) use cases. Additionally, to help organizations adopt these web-scale principles in networking, we are enhancing our portfolio by adding popular networking capabilities to the open Linux platform. These include:

Voyager code is available for early access

From the outset, Cumulus has pioneered disaggregation and openness in data center networking. We now bring this flexibility to the DCI market, where service providers, telcos and large enterprises have been burdened with un-innovative, high cost options. While the demand and the bandwidth for cloud solutions and Internet services is becoming a more tangible challenge every day, the current proprietary solutions are extremely expensive, inflexible and difficult to scale. The key to providing efficient bandwidth and cloud services across the globe is by creating open and untethered optical and IP solutions. Disaggregation with optical whitebox alternatives allows operators to be much more flexible than they were with monolithic, black-box solutions.

Voyager is the DCI industry’s first open packet-optical transport platform, combining dense wavelength division multiplexing (DWDM) technology with both switching and routing functionality. This platform is based on an open “white box” architecture that disaggregates the hardware and software. By combining the speed and long distance connectivity of DWDM optics with IP packet technologies for Layer 2 and Layer 3 networking protocols and features, Voyager offers a data center interconnect solution for metro/long-haul that is both flexible and extremely cost effective.

As announced earlier, Voyager code is now ready for early access via our partnership with ADVA. Already, organizations such as Vodafone, NYSERNet, Internet2, GRnet and CESNET have expressed interest in this innovative offering, and are testing it out. The use cases for Voyager are both long and impressive. Stay tuned as we continue with our partners to bring this innovation fully to market.

Policy Based Routing

Policy Based Routing (PBR) is designed to selectively route traffic based upon customer needs, e.g. to provide network services like DDoS protection. It simply reroutes where traffic is going based upon certain defined policies. Most routing systems and protocols route traffic only based on the destination of the traffic. But today, most environments require more than just a “what’s the destination?” approach.

With PBR, a network engineer can dictate the traffic routing behavior based on policies that change the routing behavior of specific traffic,so that you can override the routing table and influence where the traffic goes. For example, you can use PBR to help you reach the best bandwidth utilization for business-critical applications, isolate traffic for inspection or analysis or manually load balance outbound traffic.

Policy Based Routing is applied to incoming packets. All packets received on a PBR-enabled interface pass through enhanced packet filters that determine rules and specify where to forward the packets.

A couple of great use cases for PBR are around mitigating DDOS attacks and preventing backup congestion. For example, you can redirect traffic of a host or set of hosts currently under attack to a security appliance for further inspection. PBR can also be used for redirecting known large flows like a database backup over a backup path to prevent congestion on the main path.

VRF route leaking

Often, networking teams want their VRF table and their global routing table to communicate with each other in a single router.

Virtual Routing and Forwarding (VRF) allows multiple routing table instances to exist in a router or switch and work simultaneously. This increases functionality and permits multiple network paths without the need for multiple switches. Because traffic is automatically segregated, VRF also increases network security and can eliminate the need for complex security policies. This is especially important in multi-tenant networks, so they can have different clients with overlapping IP address schemes sharing the same system.

But as with the nature of anything, there are usually exceptions, requiring customers on different interfaces to share a service. As a result, in order be able to share routes or services between VRFs, at times the need arises to leak the routes between different VRFs.

This “leaking of routes” feature is now part of Cumulus Linux 3.6, where destinations in one VRF are reachable from another VRF. For example, organizations may want to leak public IP addresses to a separate VRF so that they can be handled by a different router than LAN addresses. Another example is when you want to make a service, such as a firewall, available to multiple VRFs, or to enable routing to external networks (or the Internet) for multiple VRFs where the external network itself is reachable through a specific VRF.

New 25 GB options on the Cumulus Express, Dell, Edgecore, DNI and Broadcom Tomahawk+ switches

Choice, choice and more CHOICE! We are all about giving our customers options and choice to fit ‘Any Size Data Centers.’ Whatever size switch you need, we got it. You want 25GB down and 100GB up? We can help you. Or if you want 10GB down and 40GB up, there’s a whitebox switch for you. Or 10GB down, 100GB up, no probs. We have solutions for each scenario.

This wide variety of supported whitebox switches, over 70+ supported options, is a unique and defining characteristic of Cumulus. No other whitebox NOS vendor comes close to this number of options for support. Plus, you can get all the benefits of disaggregation, like cost savings, cabling and optics choice, unification through using the Linux language across the data center, and more when compared with the larger legacy vendors.

More goodies…

The above features are just the larger features in this release. There are actually more features, such as PTP and GRE Tunneling for Mellanox boxes. To see the list, check out the release notes.