The pilot episode of Software Gone Wild podcast featuring Snabb Switch created plenty of additional queries (and thousands of downloads) – it was obviously time for another deep dive episode discussing the intricate innards of this interesting virtual switch.
- Is SnabbSwitch a framework or a product that I could download and use?
- How do I configure it? How do I change the configuration?
- How can I implement my own forwarding logic? Do I need to program it or can I reuse existing elements?
- Can I add new networking protocols or NIC device drivers?
- How can scripting code included in a configuration file forward packets between two 10GE interfaces at line rate?
- What is the difference between kernel- and user space packet forwarding on Linux?
- Why is polling faster than interrupt-based packet processing?
- How do you push 200 Gbps out of a server with 10% CPU load?
- What’s the difference between Snabb Switch and Intel DPDK?
- How does Snabb Switch integrate with OpenStack?
- How can you use SR-IOV to get blazingly fast I/O performance in virtualized environments?
Other interesting tidbits covered in the podcast:
- From the hardware perspective, you should be able to push half a Tbps through an Intel server, hilariously low forwarding rates we’re seeing on some products today are a software problem;
- Did you know that Snabb Switch uses the same technology as CloudFlare in its DoS prevention tools?
- Why it makes sense to focus on a very specific hardware platform instead of portability to all possible hardware out there.