Because I'm currently sitting on a concrete floor dicking around with 6500 line cards.
The 6700 series line cards were actually modular, in that the forwarding engine on each line card was a daughter card that you could replace!
The CFC module is the chump change "punt everything to the supervisor" card that limits the whole box to the capabilities of the supervisor.
So like... 20 million packets per second?
Only.
But God damn, that's a lot of pins on the FOUR connectors!
It even has little cute ejection levers because the extraction force is so high.
Conversely, you have the DFC3BXL, or "distributed forwarding card" (or something. W/e) which fits in the same place, and is hella bigger.
The DFC3 modules are essentially all the routing hardware of the PFC3 policy engine in the sup720, but distributed to each line card so your routing capacity goes UP as you add line cards, not get spread thinner like with the CFC
Interestingly, the DFC3 modules only use two of the connectors! Not just punting routing decisions to the supervisor over the backplane bus; it can make those decisions locally!
So now this 4x10Gbps line card has a CFC (centralized forwarding card?) So it's only as good as the supervisor you have installed with it.
Why, you might ask, am I downgrading a 10G line card?
I found a poor sucker who's willing to take my 6506, but I don't want to give away my DFC3BXL. They'll have to settle for only 20Mpps across the whole switch.
It would appear my cabinet still has the power budget left to power up two 6500s at once. 🤔
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Active Optics Cables, or AOCs, are essentially patch cables desired to plug directly into *SFP cages.
They're just like DACs, or Direct Attach Cables, but with more length options and the cable routing advantages of fiber.
The nice thing about AOCs over using SR4 optics is that it isn't connectorized, so you don't need to assemble three parts to make a link, and "cleaning fiber ferrules" is a ball game you don't need to get into.
This is a 400G OSFP AOC.
400Gbps
Octal Small Form factor Pluggable
Active Optical Cable
And let us all pause for a moment to admire how awesome OSFPs are and how they look space age af.
So let's say you have a nice server with a set of 8x3.5" disk bays in the front of it that you want to use for storage.
What you typically don't want to do is also install your OS on the same disk array.
You could burn one or two of the disk bays for the OS, but that sucks, and if you're doing something like ESXi or Green as, you don't need that much space.
But servers will have USB ports inside the case to boot off a flash drive! Literally just boot off USB forever!
But USB really sucks if your OS isn't mostly read-only, so your next option is to use one of the SATA ports inside the case.
Some server cases will sneak a 2.5" bay in the back by the IO panel.
1) Sit on the floor and enjoy this pineapple cider.
2) come up with the dumbest way possible to load test an authoritative DNS server.
Some of you might ask "but why not load test your DNS server like any... normal(?) person?"
And to that I say, "my bottle opener is a wrestler and can kick your ass"
We have these tools at our disposal:
* A potato laptop I pulled out of the E-waste.
* Some Ethernet cables, which I pulled out of the e-waste
* An Arista 7050T-64, which I pulled out of the e-waste
* My DNS server DUT, guess where I got it
* My wits
Tonight's #hamr hot-take: Your weekly nets are a goddamn waste of time if the only thing you're doing is giving your net control practice blindly reading a script and your members practice keying up the radio at the right moment and saying their callsign.
You should have two net control scripts: one that's a full script to reduce anxiety for first time net controls, and a second one that's a rough handwritten outline of all the topics they need to make sure to hit during the net.
Want to know a sure sign that a group is missing the point of a radio net and wasting everyone's time?
When they ask for check-ins, do they only ask for check-ins once, then move on with the rest of the script?
What about anyone they missed in the pileup? Do they not matter?
I've been technically assisting my father on a project where he needs to send an email to a few hundred people.
Previously, he was going it from the Gmail GUI 75 people at a time, copy-paste into bcc in the compose window, but now that I have SMTP infra, why not a for loop? 🧵
Relatively simple problem once you have the email infrastructure up, right?
cat mailinglist | while read EMAILADDR; do cat ./emailbody.txt | mail -s "Subject" -a "From: me@example" $EMAILADDR; done
So I was helping him understand the headers and how best to invoke the mail cmd
Quick and easy bash one-liner when we were ready to launch though, right?
Wrong. My father came back today with his 250 line bash script he wrote for sending this bulk email blast.