Luke Gorrie

Thanks for commenting. Have to clarify this. RISC-V is an inspiration in that it defines a simple and stable hardware-software interface. This means that people can implement new hardware with...

I talked more about this topic at FOSDEM on the weekend. Just a 10-minute talk in the context of ConnectX. https://youtu.be/ZigsHkxMHPo

@daveshah1 Great to have you here! :-) I reckon the typical Snabb user, often working in an ISP or similar environment, would go for the 1xSFP+ option. SFP+ because networks...

Just to be concrete. Some NICs like the Intel 82599 have lots of PCIe headroom and use PCIe protocols efficiently. It's really hard to make an 82599 drop a packet...

@daveshah1 One more idea kicked around on the Snabb slack channel is a PCIe3-to-PCIe2 bridge IC. Then we could have Nx10G each with a dedicated ECP5 and PHY all sharing...

@sarsanaee log scale?

This is a cool slide :). So we can state the EasyNIC goal as being to have y=500 instead of y=0.3624x+5781. That's our marketing message taken care of! ;-)

@emmericp Have you looked at total lines of code per vendor? (Are there any vendors who support 10G-100G with a small amount of driver code covering all their current models...

@mlilja01 Good points! Having the hardware de-block is interesting. I would really like to support operating at line-rate 100G on a single queue. This would mean the hardware needs to...

> The drawback of a block of packets is that it is not very protocol stack friendly. Yes. I see this as a "with great power comes great responsibility" situation....