Hedge 141: Improving WAN Router Performance

Wide area networks in large-scale cores tend to be performance choke-points—partially because of differentials between the traffic they’re receiving from data center fabrics, campuses, and other sources, and the availability of outbound bandwidth, and partially because these routers tend to be a focal point for policy implementation. Rachee Singh joins Tom Ammon, Jeff Tantsura, and Russ White to discuss “Shoofly, a tool for provisioning wide-area backbones that bypasses routers by keeping traffic in the optical domain for as long as possible.”

Hedge August Update

Monthly update for the Hedge podcast.

Hedge 140: Aftab S and RIR Policies

Regional Internet Registries (RIRs) assign and manage numbered Internet resources like IPv4 address space, IPv6 address, and AS numbers. If you ever try to get address space or an AS number, though, it might seem like the policies the RIRs use to determine what kin and scale of resources you can get are a bit arbitrary (or even, perhaps, odd). Aftab Siddiqui joins Russ White and Tom Ammon to explain how and why these policies are set the way they are.

Hedge 139: Open Source Supply Chain Security

There is a rising concern about the security of open source projects—particularly in terms of open source software supply chain. Alistair Woodman, who works closely with multiple open source software projects, joins Tom and Russ to discuss the reality of securing open source projects. The final answer? Essentially, buyer—or in the case of open source software, user—beware.

Hedge 138: The Robustness Principle

Most network engineers take it as a “given” that the robustness principle is the “right way” to build protocols and networks—”be conservative in what you send, and liberal in what you receive.” The idea behind the robustness principle is that implementations should implement specifications as accurately as possible, but they should also accept malformed and otherwise erroneous data, process the best they can, and drop the bits they cannot process. This should allow the network to operate correctly in the face of defects and other failures. A recent draft, draft-iab-protocol-maintenance/, challenges the assumptions behind the robustness principle. Join Tom and Russ as they discuss the robustness principle and its potential problems.

Hedge 137: Old FRR Defects

Zero-day defects exist in every projects, whether they are open or closed source. John Fraizer and Alistair Woodman join Tom Ammon and Russ White to discuss an old defect John found in the FRR code, the history of this defect, and the problems inherent in finding and resolving defects in large, diverse code bases.

Hedge 136: The IPv6 ULA Mess

IPv6’s designers built the concept of Unique Local Addresses, or ULAs, into the addressing architecture to make network address translation unnecessary for IPv6 deployments. As with many other plans of mice and men, however, the unintended consequences of what is a good idea tend to get in the way. Nick Buraglio joing Eyvonne Sharp, Tom Ammon, and Russ White to discuss the many problems of IPv6 ULA, why it isn’t practical in most network deployments, and the larger question of how standards bodies sometimes fail to consider the unintended consequences of a good idea.

Hedge 135: Simon Sharwood, China, and IPv6

Over the last several years various Chinese actors (telecom operators and vendors) have been pushing for modifications to IPv6 to support real-time applications and other use cases. Simon Sharwood wrote an article over at the Register on their efforts and goals. While this effort began with big IP, moved into new IP, and has been called many other names. These efforts are being put forward in various venues like the IETF, the ITU, etc. Simon Sharwood, who writes for the Register, joins Tom Ammon and Russ White to discuss these efforts.

Hedge 134: Ten Things

One of the many reasons engineers should work for a vendor, consulting company, or someone other than a single network operator at some point in their career is to develop a larger view of network operations. What are common ways of doing things? What are uncommon ways? In what ways is every network broken? Over time, if you see enough networks, you start seeing common themes and ideas. Just like history, networks might not always be the same, but the problems we all encounter often rhyme. Ken Celenza joins Tom Ammon, Eyvonne Sharp, and Russ White to discuss these common traits—ten things I know about your network.

Hedge 133: Brooks Westfield and Multifactor Testing

Multi-factor testing is one of the most important jobs a vendor takes on—and one of the most underrated. Testing across all possible configurations and use cases is nearly impossible. Brooks Westbrook joins Tom Ammon and Russ White on this episode of the Hedge to talk about the complexity of multi-factor testing and some of the consequences of that complexity.