Wednesday, May 23, 2012

P2P May Become The New Cloud, New Mobile Network

Green Pirate recently spoke with Dilum Bandara of Colorado State University’s Computer Networking Research Laboratory (CNRL) where research is being done to discover new ways in which the power of P2P technology can be leveraged to improve many existing technologies. This one is for futurist nerds, anonymous ciphernauts and anyone else who loves seeing distributed P2P technologies implemented in new ways. CNRL is currently seeking feedback from BitTorrent users as well as developers, researchers and those in various industries which may benefit from CP2P.

“How cool would it be if Open Office is online like Google Docs? Though open-source community behind Open Office will be eager to develop an online version, no one is willing to invest in a large datacenter. What if each of us can contribute residual computing, storage, and bandwidth resources in our computing devices to build a globally-distributed virtual datacenter, that can not only run Open Office but also many other FOSS tools?”

“CP2P [Collaborative Peer to Peer] system is the heart of such a community-based, cloud computing system that interconnects underutilized computing resources from homes/businesses.”

According to Bandara, there are some problems with the current cloud model that could be solved by moving to a decentralized model.

“P2P clouds will be able to target issues such as centralized data, privacy, proprietary applications, and cascading failures in commercial clouds. CP2P enables a set of community members to contribute their storage and bandwidth to store (encrypted) user files, and a second set of members may contribute their computing power for backend processing, while a third group of resources may enforce security and privacy.”

For the many people who are now faced with having their data destroyed by what is a foreign government to a portion of them, a decentralized solution may not seem like a bad idea.

Researched is also being carried out to find ways of improving the way data is transferred over mobile devices to allow for a powerful means of decentralized mobile communications.

“Mobile Social Networks – Value of social networks can be enhanced by allowing users to share diverse resources available in their mobile devices. For example, a person with a basic mobile phone could connect to a friend’s smart phone with GPS capability to locate a nearby ATM. In another example, a group sharing their holiday experiences in a coffee shop could use one of the member’s projection phone to show pictures from others’ mobiles/tablets or stream videos from their home servers.”

“Moreover, in large social gathering such as carnivals, sports events, or political rallies users’ mobile devices can be used to share hot deals, comments, videos, or vote for a certain resolution without relying on a network infrastructure. CP2P is useful in these applications to aggregate and connect groups of mobile resources based on their users’ social relationships.”

Already, weather monitoring networks which have found decentralized CP2P technology to be most appropriate have been created.

“DCAS – CASA is an emerging dense network of weather radars that collaborate in real time to detect hazardous atmospheric conditions such as tornados and thunderstorms. CP2P-based data fusion provides an attractive implementation choice for CASA real-time radar data fusion, where data is constantly being generated, processed, and pushed and pulled among groups of heterogeneous radars, storage, and processing nodes. The group of radars, processing, and storage elements involved in tracking a particular weather event continues to change as the weather event migrates in both time and space. Moreover, certain rare but severe weather events require specific metrological algorithms and more computation, storage, and bandwidth resources to track and forecast/nowcast about the behavior of those weather events.”

Fortunately, a CP2P model allows a community to offer up its resources in order to meet these needs and receive early warnings of severe weather patterns such as tornadoes.

“A CP2P system can exploit the temporal and spatial diversity of weather events to dynamically aggregate underutilized resources from anywhere in the system to enhance the resource utilization and reduce cost. Similarly, crowd sourced, community-based weather monitoring systems such as can benefit from CP2P. Armature methodologists may provide data from their home-based weather stations, while another set of community members may provide computing power to process those data to predict weather, and a third group may provide bandwidth and storage to enable public access those weather predictions.”

Bandara explains that CP2P is a common solution for multiple applications.

“P2P clouds, mobile social networks, and DCAS depend on some form of resource collaboration and strong social networks. These resources are characterized by multiple static and dynamic attributes, e.g., CPU speed, free CPU, memory, bandwidth, operating system, and a list of installed applications/middleware. These multi-attribute resources need to be combined in a timely manner to meet the performance and quality of service requirements of CP2P applications. Yet, it’s nontrivial to discover, aggregate, and utilize heterogeneous and dynamic resources that are distributed.”

“Moreover, characteristics and preferences of user communities need to be understood to better utilize resources provided by users while ensuring their incentives, security, and privacy preferences are not violated. Our goal at Computer Networking Research Laboratory (CNRL) of the Colorado State University, USA is to recognize these foreseen challenges, formalize the problems, understand the resource, user, and community characteristics, and develop fundamental building blocks that other developers, industry, and researchers can build upon.”

*I left every word of Bandara's description as I received it, so if you are not up for reading the tech talk, I suggest viewing the GREEN text in the original version which summarizes each section.

CNRL has been steadily making progress with the development of CP2P and is currently researching ways to create better decentralized search strategies.

“To date, we have shown that BitTorrent communities are not isolated, proposed a community-aware caching solution for DHT-based solutions such as BitTorrent, analyzed the characteristics of resources and queries from real-world systems, developed a tool to generate synthetics traces of resources and queries for large-scale performance study of CP2P solutions, analyzed performance of existing resource discovery schemes and illustrated why they fail under real workloads, and developed a solution for large-scale and multi-attribute resource discovery.”

CP2P researchers are currently reaching out to the BitTorrent community for feedback about how search is commonly performed today. The anonymous, eight question survey has been extend until April 8th.

“Our current survey focuses on user access of multiple BitTorrent search engines/sites with the objective of developing better community-based search strategies. Lots of work needs to be done to realize the full potential of CP2P, and we invite other developers, industry, and researchers to contribute to this vast and emerging field.”