In reply to a post by mrnelster:

So there we have it.

Length increase = noise increase = instability = Interleaving requirement = increased latency.

I am sure you will find an exception to the rule or die trying, but I'm happy with this explanation thank you.

That equation gives a reason for increased latency on interleave versus fastpath, but doesn't explain your massive swings in latency which you are experiencing. These are more likely to be a line fault.

The swings in latency were caused by fast path when it was temporarily enabled.

That's why I said fast path wasn't really an option on my line.

That's the point you dismissed originally.

In reply to a post by BatBoy:

That equation gives a reason for increased latency on interleave versus fastpath, but doesn't explain your massive swings in latency which you are experiencing. These are more likely to be a line fault.

Line faults don't cause changes in latency.

Either the packet gets through and you get a latency determined by the route taken and any congestion on it, or it doesn't and you get a lost packet.

In reply to a post by mrnelster:

The swings in latency were caused by fast path when it was temporarily enabled.

So what do you call the "pausing" if not swings in latency?

In reply to a post by BatBoy:

So what do you call the "pausing" if not swings in latency?

If there is a line fault, most likely packet loss requiring a re-send.

Absolutely nothing to do with latency.

Edited by billford (Fri 11-Mar-11 00:05:01)

In reply to a post by billford:

In reply to a post by BatBoy:

So what do you call the "pausing" if not swings in latency?

If there is a line fault, most likely packet loss requiring a re-send.

Absolutely nothing to do with latency.

When you're in a game waiting for the next packet, and it has to be resent, you will experience an increase in latency.

In reply to a post by BatBoy:

When you're in a game waiting for the next packet, and it has to be resent, you will experience an increase in latency.

Nope, it's a lost packet, nothing to do with latency.

Latency is defined as the time taken for a packet to make the round trip from the send point to the destination and back again. If it doesn't make it then it doesn't have a latency time- it's a lost packet.

One day you should learn to admit when you're wrong- it would do your soul good, save the rest of us a lot of irritation and, in time, possibly lead to your opinions being treated with some respect.

In reply to a post by billford:

In reply to a post by BatBoy:

When you're in a game waiting for the next packet, and it has to be resent, you will experience an increase in latency.

Nope, it's a lost packet, nothing to do with latency.

Unfortunately, here we are talking about gaming and the fact is, the game pauses.

We are not talking about network latency.

In reply to a post by BatBoy:

In reply to a post by mrnelster:

The swings in latency were caused by fast path when it was temporarily enabled.

So what do you call the "pausing" if not swings in latency?

As I went on to explain in the thread, The pausing was caused when BT swapped to fast path.

I had the line monitored remotely and a graph for upload and download produced. The consensus came down to packet loss and resending of data, understandable on a long line with fast path enabled.

In reply to a post by BatBoy:

Unfortunately, here we are talking about gaming and the fact is, the game pauses.

I'm not denying that, but it's a pause due to packet loss, not latency.

You could briefly lose PPP to your ISP, but it's not latency.

Someone could pull the plug out for a few seconds, but it's not latency.

It's packet loss.

You are demonstrating that mrnelster was perfectly correct- resorting to semantics. As usual when you're wrong.

Edited by billford (Fri 11-Mar-11 00:34:06)