Ants version 1.0.0 (finaly)

[tsafa1]

Ants is finaly starting to mature a little bit. Most major issues have been addressed. Supernodes working well so far, scaling issues have been corrected, optional direct connections in place, we moved to a more stable IRC channel with this version , irc.accessirc.net

speed is very good if sources stay on line. i often hit down speeds up to 70 kB/s and up speeds up to 20 kB/s

We will test this for a day before Gwren makes an installer and puts it on the source forge. so run it and post you stats.

get here:
http://www.myjavaserver.com/~gwren/Ants/antsp2p.jnlp

keep an eye out for any crashes after long up times

[Shareaza2005]

now that it's finished, let's get one of those names that were thought up for Ants, in here.

[mp3MaStA88]

what is different in this release?
and how have capabilities changed?

[Abyss00]

Ants 1.0.0
Compressed: 5% DL 4% UL
Total Traffic: 420mb DL 482mb UL
Own Traffic 0mb DL 70mb UL
Clock: 7:45
Threads: 90

[Rajarius]

So what are there, 50 people running Ants?

[Undercurrent]

Things seem to be working very well in the latest version. Ive also been getting downloads peaking at 115Kbs, however, they dont seem to be continuous. The downloads will stop every now and then. For me, this isnt a huge problem. I dont care if it takes a day to download an album or movie as long as I can get it annonymously!!! Right now i think the biggest problem facing Ants are the problems that come with any new network, lack of content!!! I know i can get what i want with Ants, if it is on the network. However, there lies the problem, we have a low user base. I know its early on in the life of the network, and only time will tell if Ants is will gather the user base need for it to truley be on the level of 2nd generatoin p2p applications. Hopefully we can get people to use Ants. To me this is the biggest issue facing Ants. Most everything else is in order. How do we convince people to sacrifice download speeds for anonymity? Which brings on the question of, Does anyone Care?

[Abyss00]

So what are there, 50 people running Ants?
Ants is still in beta testing the point is not to get a lot of users, the point is to make the program better and work out the bugs.


IP adresses are ashown in non-anon modus
Where are you seeing the IP addresses?


Biggest problem I see with this version is the high CPU usage.

[tsafa1]

Gw found a few minor bugs and will clear up today. He found somethng that was limiting traffic too and will fix. Stats also a bit off cause he found that they do not include direct connect traffic. He will fix that too.

[Abyss00]

He found somethng that was limiting traffic too and will fix.
Doesn't seem to be fixed as of 1.0.0 - 0.5.3 :(

[xrobot]

Oh... finally we can see Ants 1.0 in 2005.

[tsafa1]

version 1.0.0 portocol 5.4 looks good. I have good uploads and downloads. Plus i see more super nodes.

[Abyss00]

Yeah, a lot better then portocol 5.3

[xrobot]

we have a low user base. How do we convince people to sacrifice download speeds for anonymity? Which brings on the question of, Does anyone Care?
Most likely the best chance for Ants to grow up had been passed. Because "time" just tell us that more people don't want to sacrifice download speed, and they all still use Bittorrent and other existing network with large user base.

On the internet, everything news spread fast. I mean if Ants is demanded, it should already popular among p2p users even it is buggy and in beta stage. But it don't come up like that. As the case of the Qnext program, it even promoted on the official SUN jave page(www.jave.com), so what?

Ants developers MUST give speed at least comparable to existing network. Also, support for large file(in GB) have to be implement better. Hashing a file all the day is terrible!

[Abyss00]

From Freenet we learn that there will always be a demand for anonymous programs no matter how slow they are. Although it is not on the same level as popular programs.

From Winny we learn that as laws become more anti-p2p, then anonymous programs become more popular.

From BitTorrent we learn that there is a large group of people the demand speed above all else and that law suits only serve to make it more widely known and therefore more popular.

From Kazaa we learn that there is a large group of people out there as dumb as rocks.

Ants' fate is more linked to the outcome of the Supreme Court case about p2p then it is about speed.

[xrobot]

Also, too emphasis on Ants is a Java program is no good. Although the java web start installation seem advance and convenient for development, for the general end-users(who don't notice java or c++) will associate the fear of forced installing spy-ware on the net.

I suggest the Ants developers should focus on general downloadable installer, instead of java web start, as the main method to spead the software.(even remove it for developer or advance user only).

Azureus is one of the very successful case in suppressing the effect of the critism on Java. It promote the software feature, plug-in in the main context. JRE is just need to mentioned when you want to install Azureus, as a system requirement.

[Abyss00]

What is the current stable version now? For people that don't want to update all the time.

[tsafa1]

comming soon....

[Hornet]

In ANts P2P beta 1.0.1 protocol 0.6.5.

Can you or Gwren explain: why when in anonymous mode and you are downloading a file from a user in DC mode (so you can see their ip) and coincidentally they are your neighbour - why do downloads STILL don't start and even if they do start only go for about half a second before stopping.

You are downloading from a neighbour with free slots YET STILL CANNOT download :(

I am downloading from my neighbour so where is the hold up?

I know the source is my neighbour as I can see their ip, right, and I can see my neighbour's ip and they are the same - so no intermediate proxy for file transfer - yet still does not work.

Does network choose not to download or to download very slowly?

I know the ip of source as source is non anonymous - therefore I can see the source is my neighbour - yet I still cannot download although source has free slots.

Before non anonymous mode you could not see if the source was also your neighbour ( source could have been 5 hops away). So this question did not come up.

Could you or Gwren explain this with a diagram for simpletons like myself or is it a secret?


Hornet :hole :playboy :hole

[ABC_thellookoflove]

use azureus or emule over i2p

[Hornet]

use azureus or emule over i2p

i2p does not work but neither does ants!

Why is it after all these updates that ants beta works worst than the alpha?

Can Tsafa or Gwren explain:

Why when in anonymous mode and you are downloading a file from a user in DC mode (so you can see their ip) and coincidentally they are your neighbour - Why do downloads STILL don't start and even if they do start only go for about half a second before stopping?

You are downloading from a neighbour with free slots YET STILL CANNOT download :(

WHY?

I am downloading from my neighbour so where is the hold up?

WHY?

I know the source is my neighbour as I can see their ip, right, and I can see my neighbour's ip and they are the same - so no intermediate proxy for file transfer - yet still does not work.

WHY?

Does network choose not to download or to download very slowly?

Am I the only one having these issues or does everyone else just give up and wait for the NEXT update which might fix this BUT NEVER DOES.

I know the ip of source as source is non anonymous - therefore I can see the source is my neighbour - yet I still cannot download although source has free slots.

Gwren should issue some documentation so we can see WHY.

Before non anonymous mode you could not see if the source was also your neighbour ( source could have been 5 hops away). So this question did not come up.

Could you or Gwren explain this with a diagram for simpletons like myself or is it a SECRET?


SO MUCH FOR OPEN SOURCE.


HO r Net :cross :cross :cross

[tsafa1]

The reason the alphas works beter then the bata's sometimes is because they use near 100% cpu power. So Gwren goes back an he cuts some fat. Then you see loss of performance so he needsto adjust that. After that is working its time to add a new feture. Then we go back to the begining of high cpu and we need to cut some fat and so on. You can expect this cycle to continue untill Gwren runs out of things to add. That is what it takes to develope a program. Just look how long it too gnutella to become functional from original prototype... 4 years? If there was an easier way to do it Gwren would do it or someone else would do it ahead of him.

[Hornet]

The reason the alphas works beter then the bata's sometimes is because they use near 100% cpu power. So Gwren goes back an he cuts some fat. Then you see loss of performance so he needsto adjust that. After that is working its time to add a new feture. Then we go back to the begining of high cpu and we need to cut some fat and so on. You can expect this cycle to continue untill Gwren runs out of things to add. That is what it takes to develope a program. Just look how long it too gnutella to become functional from original prototype... 4 years? If there was an easier way to do it Gwren would do it or someone else would do it ahead of him.

OK, So you agree with me that it does not work.

However, I think it needs more than FAT trimming.

Current issues with ANts:
Direct Connection downloading (DC) does not work,
Can't download from neighbours although neighbour has free download slots,
Downloads stop after an hour and only start back up if you restart ants,
Upload speed is not calculated as most users would expect (speed is one twentieth of what it should be) and
Super node searching does not work (often when you (tsafa) are connected I cannot find any of your files).


Before release candidate, ANts needs to do two things:
NO MATTER HOW SLOWLY, it should firstly be able to search and find a file first time, everytime, if it is on the network and
NO MATTER HOW SLOWLY, secondly download that file if the node where the file is situated has free download slots.

Ha Net :mellow :mellow :mellow

[Hornet]

Why when in anonymous mode and you are downloading a file from a user in DC mode (so you can see their ip) and coincidentally they are your neighbour - Why do downloads STILL don't start and even if they do start only go for about half a second before stopping?

You are downloading from a neighbour with free slots YET STILL CANNOT download :(

WHY?

I am downloading from my neighbour so where is the hold up?

WHY?

I know the source is my neighbour as I can see their ip, right, and I can see my neighbour's ip and they are the same - so no intermediate proxy for file transfer - yet still does not work.

WHY?
Explanation based on Gwren's notes:

I think the explanation is the New Routing System
http://www.myjavaserver.com/~gwren/Ants/New%20routing%20system.pdf

In this, the route chosen by a message is not the shortest or fastest but according to connection load, “least charged connection is chosen for each message “.

So if your neighbour is the source of the file but there are routes to the file through other neighbours then the fastest or shortest route is not used. Instead the route through the neighbour with the fastest connection is used. So it does not matter that your neighbour is the source of the file. The shortest route is not used. The reason given for this is to make the network automatically load balance.

However, a bye-product of this method is that very often you cannot download directly from a neighbour. Instead the file is routed through several hops, often unsuccessfully to your neighbour.

I believe this is the wrong way to do it. It is not simple. It is counterintuitive. It creates unnecessary long paths.

Also, it does not load balance as connection speed is not a reliable guide of network congestion.

A relatively slow computer connected to a very fast T3 connection could be more congested than a very fast computer connected to an ADSL connection!

Moreover, there is already a way to load balance the Propagation System.
http://sourceforge.net/mailarchive/message.php?msg_id=10881813

“Basically when a node disconnects from another node, they both send to every remaining neighbour a message with a list of all the peers (IDs) now unreachable (because of the disconnection), this message is processed and it is also forwarded in a linear way to all the net, to let peers update their routing tables.

To be precise the message is not trivially broadcasted, there is a list of the unreachable peers in it, after the message is processed by a node this list is usually made shorter, because there could be alternative way to a peer.

So actually it can be unreachable from a peer but not from another one.

So step by step the list get shorter and shorter, when the list become empty (usually few steps because there are MANY different routes to a destination) the message is discarded.

This is a smart way to propagate the information in a "light way".”

This Propagation System should be extended to cover message time-outs, as these are due to congestion.

When there is such a time-out, the routing table should be updated and a peer-unreachable message should be sent to the appropriate neighbour.

That neighbour would then update its routing table and then forward the message to the appropriate neighbour and so on until it reached the source of the file or the message could go no further due to a time-out.

In this way congested routes that often time-out would be dropped and routing table would be kept updated.

This would also solve another issue.

Often when you first connect to ANts you get very good performance, fast downloads and many sources.

However, after about 45 minutes, performance tails off and downloads stop and sources disappear.

You then have to either close and restart or stop and start Ants to restart downloads.

A reason for this peculiar behaviour maybe because of routing table getting old and not reflecting congested paths. Restarting Ants updates routing tables.

An extended propagation system would solve this issue.

What do you think?


Ho R net :blah :blah :blah

[ABC_thellookoflove]

we needs for ants a working search mechanism and as well a seperation of core and gui.
Many projects are waiting for ants, downloadsites as well als downbattlehill gaim plugin.
i2phex will come soon and as well azureus has workign i2p network integrated, still nobody knows howto, but in 4 months all is done!
Dunno why there is no progress. Time is not the question.

[CactusChris]

Hi ABC
I agree with you that the searches are not working - often i get zero results on queries that should return many results.
I have also noticed that CPU usage goes up with time, and eventually heads to near 100%. The amount of data being proxied is high, but download ratios are now below 1% of throughput, this may be down to my selections, but I think it is related to the search issues.
A further small issue is thet when minimising and then 'unminimising' the initial picture is not in the right place, and clicking on it often gives no activity, only clicking where it really is (a bit above) resets the box where it should be -probably a java thing...

BR
Chris

[ABC_thellookoflove]

Ants version 1.0.0 (finaly) has 10 users.

[tsafa1]

Things are gona be quiet untill Gwren gets back from Ireland or untill some other programmer starts tinkering around with the code.

[ABC_thellookoflove]

right and this is why http://www.i2phex.tk will win,
Though I like the integrated jetiants, but if it is not working, we have to forget it.