Smart Path Selection?
This thread has been locked for further replies. You can start a new thread to share your ideas or ask questions.
If I have three of these set up, and the connectivity differs between two devices compared to others, can they intelligently choose the best path to a remote resource?
Example:
- Client device sits behind adapter A
- Server device sits behind adapter C
- Connection speed from A->B is 250Mb
- Connection speed from B->C is 250Mb
- Connection speed from A->C is 150Mb
Will it intelligently choose the path from A to C via B? Or does it always opt for a direct connection?
What about path control to prevent broadcast storms?
I'm experiencing delays that I don't expect, and I'm wondering if it is because the devices are not intelligent and able to choose the "best" path on their own. And, if that's the case, is there anything I can do to influence selection of the path?
- Copy Link
- Subscribe
- Bookmark
- Report Inappropriate Content
Powerline is a one-to-one connection. Each remote adapter is linked to the Main or Host adapter, so there is no relay through other remote adapters for improved signal quality. The article below explains some of the issues with powerline that resullt in slow speeds. Also keep in mind that distance and eletrical interference also play a role in performance. Any major appliance we would recommend being connected to a surge protector. Not only do surge protectors protect your devices from surges but they also filter out the electrical noise a device makes that affect the signal as it passes down the line.
https://community.tp-link.com/us/home/forum/topic/150615
- Copy Link
- Report Inappropriate Content
Thanks for the response.
ALL of my equipment is connected to power-filtering UPS devices as surge suppresors are not terribly useful overall and do nothing to protect against power sags (brownouts) which are much more common and destructive than a spike. The problem is that PL adapters have to be direct-connected to power for optimal performance and that means that there is no way to completely isolate power spikes from wired network connections (a spike could travel through the adapter, along the CAT5, and into the connected device and could futher propagate from there).
I understand how these limitations impact the connectivity between any two devices. A was my first device and B was second. C was added afterward, and I am now suspecting that A is functioning as my "hub" device and ultimately is causing my issues. B should be the central device in my setup as the "distance" (across the power connections) is affording for better overall connectivity through the system. If A is my hub, then it means that a device connected at B has to first travel to A and THEN can travel to C to get to that same server resource, and that is absolutely non-optimal.
How to specifically designate one of these as the "hub"? I'd like to try both B and C as the hub in my setup as both A and B host clients, but B has the best overall speed for A to get to C as direct right now is not optimal because they are far apart on the breaker panel.
- Copy Link
- Report Inappropriate Content
So I discussed this further with our testing team. The Agent in office was acutally the one who what tasked with powerline. The way it was detailed is a little different than i said earlier. It is true that the adapters are a 1 to 1 connection, meaning A - C, B- C, A-B, etc. But also there is one adapter on the network degistnated as the COC "command operations center" This unit determines which unit can talk and at what time on the network. This allows for clean communication and collision advodiance. The COC device cannot be chose by the user though. It is determined by the 1st adapter added to the network. So theoretically if you disconnected all 3 adapters the 1st one to be put back in placed and power up would be the COC.
I hope that helps a bit.
- Copy Link
- Report Inappropriate Content
Thanks for the additional information.
There's still something that isn't clear, though.
You first stated that the COC is the adapter that's "added to the network" first. Then you said that powering all three off would result in the first one powered on becoming the COC. So, it seems that the first one plugged in and powered on is the COC which is easy to follow. But that leaves a question for me...
In a scenario where the COC loses power and then returns to the network, and the remaining adapters remain powered on, what happens? Is a new COC chosen when the original one goes off line? Or does the whole network die? If a new one is chosen, what happens when the original one returns into service?
I'd also like to better understand how the devices are allowed to transmit packets in a scenario where devices at location A and B -both- need to communicate with a device at location C at the same time?
- Copy Link
- Report Inappropriate Content
If the COC device leaves the network for any reason the next device added will become the COC. For the 2nd question its not possible as each adapter has what is called a timeslot although this is in miliseconds it is literally impossble for the adapters to transmist at the same exact time.
- Copy Link
- Report Inappropriate Content
More good info - thank you.
One last technical question: What happens in a scenario where there are multiple adapters and a disruption causes a partial loss of communications? Let's say we have A, B, C, and D all on the network and A is the COC.
What happens if communications between A/B and C/D is disrupted? I would expect that the A/B portion of the network would simply understand that C and D are not available. But, if C and D can still talk to each other, it seems that a new COC would be chosen.
Fast forward a bit to where full communications is restored. The previously separated halves of the network, each with their own COC, has been "reunited". What happens in terms of a COC? What happens in general?
- Copy Link
- Report Inappropriate Content
That would kind of be hard as the 4 adapters are all paired together and work on the same electical panel. For something like this to happen adapters would likely have to be removed and placed in circuit controled by a seperate panel. In that scenario the adapters will be added to the existing COC when returned to the same circuit and rejoin that circuits already existing COC.
Another example is an outlet that an adapter is connected to goes out. When that adapter is connected to a new outlet it will scan for its network if avaialble it will restablish pairing and become a link in the chain.
If an adapter somehow loses its pairing all together then simply adding it back in via the Powerline password every adapter has on the back label will restore its link to the pair.
I hope that answered the question.
- Copy Link
- Report Inappropriate Content
Yes, that is very helpful. I realize that theoretical questions are hard to give finite answers to, so I appreciate what you were able to share.
Last question - why isn't this documented and available on line? I realize that much of your hardware is intended for the home user market, but I work with networking technology and many of my peers use TP Link hardware for a variety of things. I am certain your utilization rate would go up if those that are "in the know" could better understand your technology (and possibly even suggest changes to make it better).
- Copy Link
- Report Inappropriate Content
Quite honestly power users don't typically engage with us and ask more detailed questions regading how a device works. Most of the questions for home products are how to setup or to change settings. Hence the reason why so many of our FAQs, Videos and guides revolve around those subjects.
- Copy Link
- Report Inappropriate Content
I can completely appreciate that. Home Users want Home User Relevant Information. My point was more toward the "how to attract more power users" sort of area.
At any rate - I am thankful for all of the information.
- Copy Link
- Report Inappropriate Content
Information
Helpful: 0
Views: 863
Replies: 10
Voters 0
No one has voted for it yet.