Author Topic: How's the primary side of the client going? (Read 23468 times)

achilles · « **on:** February 23, 2012, 08:27:01 am »

Any word on how progress is going on the primary protocols? What can we expect when the primary side is complete? Any chance we might see a different hashing system used? Maybe something similar to DHT?

Lagerlout666 · « **Reply #1 on:** February 23, 2012, 10:19:26 am »

The client will do everything winmx will do at release. After that who knows, but the target is to have a complete clone up and running and out the door soon.

achilles · « **Reply #2 on:** February 23, 2012, 11:28:54 am »

I hope the network stays truly decentralized because if not then I don't believe the network will last long in light of recent events. I hope there's no single point of failure in the network. If this were the case then pressure could be placed on a select few to bring Down the whole network down.

silicon_toad2000 · « **Reply #3 on:** February 23, 2012, 11:45:08 am »

Being decentralised as it is is one of the key benefits of winmx, there's plenty of other p2p networks out there with centralised aspects.

achilles · « **Reply #4 on:** February 23, 2012, 04:26:30 pm »

Networks that are not decentralized invite legal pressure to shut them down, and prosecution for those that hold the servers that keep them alive. When only a few users hold the life line of a network in their hands they will become an easy target for those I need not even mention. When your have hundreds of thousands or even millions of primaries keeping a network alive with highly secure protocols then it would be near impossible to bring the entire network down. At the lease it would be an extremely costly, and difficult thing to do. These are topics I like to discuss, but probably would be best discussed out of the view of those that are threats to our way of computing. I'm very excited, and eagerly waiting to see a highly secure client supporting a completely decentralized network! I believe the network will be much lager than it ever was before simply due to the fact it is so easy to shut down those P2P technologies that the masses are currently using today. Not only is it easy to shut them down; it is happening at an alarming rate. I would predict those certain technologies / networks will be on the endangered species list before long. I need not even mention any clients or networks by name. Especially since I have nothing against those networks and or clients. I mean no disrespect to any of them. My philosophy in life has always been give credit where credit is due. We don't all have the same preferences as one another. If we did then life would be pretty boring. My prediction is we will be seeing a whole lot of new users come to the network soon. Its also my belief that many of them will be thinking to themselves wow.. I wonder why I never new about this network before. So stay the course and I myself like many others are eagerly awaiting an even better network than we ever had before.

Joshua203 · « **Reply #5 on:** February 24, 2012, 07:25:23 am »

Quote from: achilles on February 23, 2012, 04:26:30 pm

Networks that are not decentralized invite legal pressure to shut them down, and prosecution for those that hold the servers that keep them alive. When only a few users hold the life line of a network in their hands they will become an easy target for those I need not even mention.

Well does n't that make you wonder where we would be if the peercaches were shut down for a longer period achilles?
It s the sole reason why frontcode left us hanging years ago, and we were dangling by a thread for a week (yeah only a week ... LMAO)

GhostShip · « **Reply #6 on:** February 24, 2012, 03:26:41 pm »

All networks require some form of node discovery thats how the internet works, its simply a case of which is the simplest method and in the case of MX the one that ensures no liability for the cache operator.

By ensuring caches are located globally we hope we have addressed the issue of another shutdown, frontcode had all their caches located in a single data centre in the US and thus when a highly speculative claim of liability was made it affected Frotcode directly.

RebelMX · « **Reply #7 on:** February 24, 2012, 05:42:07 pm »

Has it been considered that the program actually stores a local file of primary node info as perhaps a back up? Since the primary info is shared in 8's (or 10's) it may be useful to store a list locally for any possible time of need, maybe even a setting that is enabled by default that allows a user to turn this feature off if they wanted?

Lagerlout666 · « **Reply #8 on:** February 24, 2012, 07:59:42 pm »

As far as im aware it will do that, meaning once connected a primary client would never need to access the caches again, even after its closed down and re-opened. unless all its nodes are offline. I remember discussing it anyhow, i forget all these little bits

White Stripes · « **Reply #9 on:** February 24, 2012, 11:01:02 pm »

a feature to manually feed it ip:ports (as either secondary or primary) to get up and going again during a 'catastrophic failure' would be an interesting feature creep in future versions....

but an attack immune clone of mx3 first.... feature creeps later...

Joshua203 · « **Reply #10 on:** February 25, 2012, 09:26:34 pm »

yeah that would be a nice feature in case of unexpected problems to overcome a small period of downtime of ALL caches (if that would ever happen)

Nice screeny of the new feature BTW

Bluey_412 · « **Reply #11 on:** February 26, 2012, 12:15:35 pm »

That does indeed sound almost like the 'super primary' concept that was discussed somewhere else recently, and which i really liked. Such 'super primaries' would be a kind of primary client+cache in one package, hmm, subject to 'limited distribution'?

White Stripes · « **Reply #12 on:** February 26, 2012, 01:57:49 pm »

'super primaries' sound like an easy way to cause a lot of havoc...

Bluey_412 · « **Reply #13 on:** February 26, 2012, 02:10:28 pm »

'Limited Distribution'?

GhostShip · « **Reply #14 on:** February 26, 2012, 03:21:27 pm »

There are no plans to construct a "super primary" the two tier mechanism we have atm is one of the most optimal for peer to peer networks.

The node discovery/storage enhancements mentioned here are common sense ideas to build in redundancy against a catastrphic event, thus its prety certain all such mechanisms of this sort will find their way into the project.

achilles · « **Reply #15 on:** February 26, 2012, 03:45:12 pm »

With a well designed primary side of the client all we really need is enough users doing their share. More and more users are having broadband brought to their area. If a good percent of the users that can will run as primaries then the network should be rock solid. Some users such as myself will make good dependable primaries since I have good bandwidth, and I leave my Icore 7 3.2 ghz machines running for weeks or months before I reboot. I don't believe there will be a need for super primaries with a well designed client. I'm not really sure how the network would be crippled or shutdown to ever have to use them. I have nothing against using them though unless it creates a legal liability for those running them. If its just a limited distribution then maybe it would lower the liability involved. I can't really say. The more decentralized the network is the more resistant it will be to attacks physically, and legally. Bluey, could you explain more how they would be used or why they would be needed? No disrespect intended here. I'm learning more everyday about P2P protocols. Wish I was a coder to be of more help.

White Stripes · « **Reply #16 on:** February 26, 2012, 08:44:59 pm »

Quote from: Bluey_412 on February 26, 2012, 02:10:28 pm

'Limited Distribution'?

like the source code to eagles patch and the cache server?

RebelMX · « **Reply #17 on:** February 26, 2012, 11:11:28 pm »

Technically all Primaries are "Super Primaries" in that they run as primary nodes and also distribute to others a list of other primary nodes (similar to caches).

Therefore all is needed is a file with a list of primary nodes, saved locally and also shareable over the network or on support sites in case of a large scale fail.

GhostShip · « **Reply #18 on:** February 26, 2012, 11:24:57 pm »

Your concept of a super primary is not the same as my own Richy, I cant speak for anyone else here but a super primary in my book would be one able to handle Gnutella levels of primary network nodes (32)

@ Stripes

I don't find it at all funny that one person was trusted with both srcs and sold the community out, the distribution of those src codes was pretty small and for the dates on the files even smaller in fact the creator of one of the files and Pri, that is after all how I was able to know who it was, there existed newer and older versions of the cache src code but the specific one Pri had was never distributed to any of the other cache operators as it was under test at the time and a change was added before the next release thus confirming it was him. The network attackers made good usage of the src codes to find ways to hurt the communities volunteer peer cache operators, we lost 3 thanks to such underhand attacks.

When folks are ready to share material that they have promised will never be shared outside of the tight community here and then make the mistake of handing it to a disgruntled ex community developer we here are left to clean up the mess as usual.

I have learned some users will lie and lie again rather than admit they messed up, and the impact of such folks poor actions unfortunately hurts the rest of the community, the new project src is not shared to anyone except the developers and thus others have lost out on learning to help create usable code for the community and having a hand in the creation of its future, we have all lost time, opportunity, and tens of thousands of users thanks to such folks, I hope they hang their heads in shame.

White Stripes · « **Reply #19 on:** February 27, 2012, 12:19:19 am »

@GS

funny? no its not funny... i was being sarcastic...

blueys 'limited distribution' would be very limited indeed due to the small (and shrinking?) pool of the trustworthy....