This is an interesting argument that reminds me of the difference between "communication" and "connection," both at a technical level and a human level. Commenter ppietra's points are valid for many networking protocols that support both connected and unconnected messaging, where connected messaging generally infers that there is a notion of session/connection state information that is maintained by the endpoints and intermediaries that exists even when packets are not being actively sent of the wire. This contents, context, lifetime, semantics, and roles related to this state information is protocol specific. That's all fine and good.
The human side of this argument falls along the lines of the distinctions drawn between communication and connection that are the primary focus of writers, speakers, and presenters like John C. Maxwell, most notably in his book "Everyone Communicates, Few Connect." I'm not going to rehash the book here, but it is very evident that there's a whole lot of communication taking place around the topic of connections, but not very much connection is actually happening.
Your mention of sessions makes me think I know where those chuckle heads were misunderstanding basic networking lingo. Take TCP v UDP for example, which many here probably learned about at one point. One is considered connection-less, but that's simply because it doesn't do the three-way handshake before sending data, which is first reaching out, then being informed that you were heard by the recipient, and then having the initiator reach back out to the recipient to say they got their message before sending the encapsulated data. There are also additional messages (read: overhead) for verifying the data that was received yada yada yada that makes TCP a heavier protocol over UDP, but in both cases a connection has to be made, it's just that with UDP and many others don't first establish a connection prior to transmitting data. You let other protocols and often higher layers figure out what was and wasn't received and then make additional requests as needed.
But that's all higher up in the OSI model. Even if you could use all connection-less protocols (read: protocols that don't establish a prior connection before transmitting) they still need would need to be run on on a connected network at the lower layers before they can operate and then would need other layers to verify that data was received. Even something as basic as plugging in an Ethernet cable connects something. In this case it's physical connection, and that can sometimes be the problem if there is a fault in the hardware.
Each connection opens up the possibly for more complex connections but the model still works the same regards of how technologies evolve. I've spent decades designing, building, optimizing, and troubleshooting networks. I can't tell you have many times I've had to come in to resolve was I was told were unsolvable problems that turned out that they were simply digging at the wrong layer, so to speak. I can't fault those who don't work in IT for not knowing, but I'll be forever perplexed by someone in IT that can't figure out why someone's [insert app] isn't connecting without them ever trying to ping an IP, go to another website, or even check the bloody physical connection to narrow down the scope of the problem. A systematic approach has never failed me.
I preciously used the example of establishing links to satellites because there is no handshake. Usually you see the word link in this context, but a link is just a connection. First the physical layer, which will mean the radios and antennas with the correct modulation for that medium and without obstructions/interference. If these aren't congruent then there is no connection. Period. Then the data link layer, and so on. These are still connections even if we don't see a physical cable running to the satellite, without a three-way handshake between nodes, and without the satellite verifying that we're receiving data.
PS: I'm still laughing about someone saying that Bluetooth and NFC aren't a part of networking. I guess this place hasn't changed much.
I can’t remember - can the firmware of airtags be flashed? Presumably Apple could release a firmware update to eliminate/mitigate the effects of a hack.
As far as the hack itself, I’m having a hare time seeing much risk.
Comments