Not so fast. There are many scenarios where a life extending charge method could be really helpful, including cellphones and EVs. Perhaps you like to get a new cellphone after 2 years and don't think doubling the total life of the battery is worth it, but wouldn't you like it if your nearly 2 year old phone still held a charge like a nearly 1 year old phone? Don't you thing that at least for some people that might make it worthwhile to hold on to it for another year?

For EVs, one of the biggest worries is how much it will cost when it's time to swap out the batteries. Don't you think being able to put that off for 4 years might be worthwhile.

Pulsed charging won't likely require much modification to charging circuits.

New battery formulations can take years to go from proven in the lab to available to buy. A pulsed charge circuit should have a much shorter lab to street time.

Not my router, so I need to defend against the screwed up config. I choose to do that through a combination of network setup WRT routing and using VLANS to keep traffic that shouldn't be away from my maintenance net. If there should never be traffic from the uplink port to the maintenance net, just block it just in case. Defense in depth.

As a side note, that's also why I avoid sharing the host port for the BMC once a box is in production. It has been handy figuring out what's wrong when helping hands turn out to be less helpful.

Cisco has a bad habit of doing proxy arp.

If martians could come from another customer's network to mine, I have no reason to believe it couldn't go the other way.

The colo manager I contacted about it thought it was anything but normal. The 10 net should have been null routed, of course.

You may be surprised to learn that the little bitty microcontroller most BMCs are based on have significantly less computational power than a 32 core Epyc CPU does...

And that is why I would VLAN my uplink off from my management network. I don't trust your router's config...

You seem to mis-understand security. It's not belt OR suspenders, it's belt AND suspenders. AKA security in depth. I wouldn't depend on just VLAN tagging for security. I wouldn't depend on just routing and firewalling for security.

Finally, do you now or have you ever used Solarwinds?

And I do pentests. I have also seen actual cases where I was getting martians from the mis-configured upstream router in a colo. The martians probably came from another customer in the datacenter. That's how hacks spread from domain to domain though in that case I don't think anything malicious was going on.

In another incident, I found a router that had telnet open reachable through a dial-up. It wouldn't take long to guess a password given that it wasn't logging failed attempts.

Note that even if you only manage one way routing, there is potential for mischief. Anything from DOSing the BMC to a blind attack.

So I really don't care what you manage, I just hope someone else there knows better than to put all the eggs in one basket.

Just repeating yourself won't make it true.

Let's say I hack the upstream to router to actually route the 10 net. With properly configured VLANs, the BMC STILL won't see the packets. Belt and suspenders.

That would be the company's fault for creating the illusion that they would provide those things. If I lie and someone believes me, the consequences of that belief are all on me.

Also, I doubt HP actually loses money on the printer, they just don't make a lot of money without the tying. You'd be amazed how much mark-up is put on things these days. It's why you sometimes see prices so low it's almost silly buying direct from Chinese companies. Note that the much more expensive product from an American or multi-national was probably made in the same factory by the same people.

Somebody's got to pay for the expensive execs, private executive jet, and substantial campaign contributions.

VLAN can be as secure as seperate switches if done right. If the BMC's packets are tagged VLAN2 and the uplink port is VLAN10, the BMC will never see any packets from the internet even if someone manages to get an upstream router to route the 10 net. Packets from the BMC tagged VLAN2 are not going out the uplink even if the destination MAC address is the gateway. ARP won't happen anyway so the BMC can't even discover the gateway's MAC address.

Not if you configure things right. If the switch is correctly configured to tag packets incoming from it's uplink ports with the public VLAN tag (whatever vlan number you designate for that) and you put the BMCs on a DIFFERENT VLAN, they are not visible to the public.

THEN you put the BMCs on a non-routable subnet.

Then you use a well firewalled jump box with access restricted to authorized techs to allow remote access to the BMCs.

There IS a risk there that if you allow the BMCs to share the host's port rather than using a dedicated port, if one of your servers is rooted, it could be made into an un-authorized extra jump box.

Surprise, servers need to be configured correctly to avoid tears.

SM makes the BMC easily reachable on dedicated and shared port by default so you can order one in, have a remote person slap it in the rack and hook up cables and then you can set it up remotely.

I'm not a fan of using the shared port on a publicly reachable LAN since a compromised machine could be turned into an un-authorized jump box to the management VLAN, but there are times when it makes sense on a private network.

But it's not quite as good as "She sells sea shells by the seashore".

That's basic stupid engineering. All that stuff that needs calibration? Put it in the frame AROUND the big sheet of glass that chips and cracks frequently due to road hazards. Or behind the glass but not attached. If maximum view is needed, how on the edge of the roof above the big sheet of glass (really composite).

The heater coils are OK in the glass, they don't need to be calibrated, just plugged in.

Repairability and maintainability are important engineering parameters that have been largely ignored.

Yes, every business claims inflation, especially the ones whose price hikes well exceeds the inflation rate.

I can see their point about the cost of repairs. That *IS* greed driven, but it's not their greed, it's a combination of auto-maker greed and poor engineering. How about not putting crazy expensive sensors on the windshield? Perhaps protect other sensor placements a bit. Don't make entire light assemblies into thousand dollar monolithic modules. And don't charge $700 for a camera that can be replaced with an off-the-shelf part for under $100.

Yes, rounding errors are a lie. That's how a candy that's made of sugar and flavoring gets away with claiming zero grams of sugar.

Does that not sound a bit like a lie?

Also you get ludicrous serving sizes to make the numbers look better. Who eats half a cookie?