Now, we live in the age of sharing, when we must instantly tag and share that photo of our friends snorting milk out the nostril that perfectly captures how hilarious you think you and your friends are and which must be instantly shared with everyone. So realising that I had never really used a P2P network made me feel like some kind of outsider.
As I understand it, a peer to peer network involves the sharing of resources and services by direct exchange from one 'peer' to another. A peer in this sense is a computer or user who has the ability to upload and download information that can then be sent to others. The main aspect of a P2P network is that is involves direct collaboration.
Contrastingly, the main type of network involves a client/server scenario, where a server (mainframe computer or hub) is situated as a sort of middle man between each of the individual client computers. This server contains all the information that any of the other computers could possibly need. When one of these computers need some data, the server merely searches through its massive database to find what it is that is being searched for and sends it out to the computer. However, herein lies the fault with client/server networks.
Whilst this type of network is well-known, powerful, easy to maintain and is generally a very successful model (the World Wide Web functions in this nature), the fact that it is the server who contains all the information means that if it crashes or experiences failure, then the whole network is accordingly affected. This means that data cannot be accessed, a web page cannot be accessed. Indeed, you cannot connect or access anything.
Another downside is that because there is only one server at the centre of a large number of other client computers, the mass queues created more often than not, slows down the communication quite extensively.
What a P2P network enables is there to be no single point of failure. No server, no central location for all the trouble to be contained within. Instead each individual peer acts as a server containing and contributing resources such as memory, disk space and network bandwidth. Also within a P2P network the individual has the option of leaving whenever they feel like and not have to worry about the effects that might be had.
There can, however, be hybrid model P2P network applications. Whilst the hybrid is P2P in the sense that files are still be transferred from peer to peer, a central server still exists as middle man, bringing back the risk of single point of failure to that particular application. The model is still centralised.
A fully decentralised model involves no central server at all, resulting in no single point of failure, but also no central authority responsible for regulating control of the application. More often than not this leads to not being able to ensure results and having so much information present that whilst it can be searched, there is just no feasible way possible to do so.
At the end of the day though, P2P networks have the potential to be really great file sharing systems. A further development is for structured P2P networks, which guarantee the number of connections between each of the nodes in order to answer a particular search entry, although research for this has slowed down at the moment apparently. Like I would really know, being the big user/understander of P2P networks that I am.
P2P: when sharing is not necessarily caring (it's basically just take, take, take).
Image 1 available: http://www.techsoup.org/learningcenter/networks/page4774.cfm
Image 2 available: http://en.wikipedia.org/wiki/Peer-to-peer
No comments:
Post a Comment