Going Peer to Peer

It may be a matter of coincidence but both topics we have chosen to discuss in this month’s issue of the magazine have been over-hyped. While XML has been around for around four years, you can track the history of peer-to-peer computing to maybe over a decade-and-a-half. But with both technologies becoming hot topics of discussion after so many years, it is clear that these are technologies which a modern day developer, an IS chief or even a CTO can never ignore.

Peer-to-peer technology is moving so fast and is influencing so many that even the biggest consultants and IT managers aren’t sure what vendors mean when the P2P label is used. But learning what P2P can do for your company is perhaps worth knowing.

A Bit of History

We’re all familiar with workgroup or SOHO LANs that have users sharing printers and disk space with each other instead of through a single managed server. What is new, however, is purposefully opening up the concept across the Internet—something that has come to be termed “P2P”. In fact, it existed even before the Client Server networks of the mid-nineties where the clients used different sharing technologies to share disk space and even printers. P2P networking was considered very crude as we moved to the Client Server architecture and, in fact, almost died with Novell Netware 3.x and the dedicated Servers entering the corporate LAN space.

Into the new millennium, the P2P concept has been made immensely popular by companies like Napster and Gnutella, which offer freeware applications that let users share digital music and other files with each other. And while Napster’s business model (no revenue stream and tremendous litigation expenses) is questionable, it does demonstrate technology that can be applied in useful and, yes, even legal ways. In other words, P2P can be used to share files and information—from inside and outside an organization—more efficiently.

In fact, networking technology historians will note that all of a sudden the very paradigm of network computing is coming a full circle. In the 1980s, it started with peer-to-peer, then it moved to client-server. With the internet, the servers became bigger and later the clients smaller and thinner. Now we are back to talking peer-to-peer. (The future could be Rich Clients)

Some P2P visionaries see direct peer-to-peer communication as the next phase in computing —as a model that has the potential to replace the server-based Internet model. While no one can say that P2P will make the Internet and client-server disappear in one fell swoop, it seems likely that at least some mission-critical apps will step off of the server. More importantly, some apps probably should, and it’s probably time to figure out which ones those are in your enterprise.

Through P2P, companies can solve critical problems in networking, information management, communal work and communication, community-building and file sharing. Many of these tasks can’t be performed as well or as quickly through the traditional client-server model.

And P2P technology offers new benefits of its own. For example, according to P2P developer BuzzPad, serverless, direct P2P communication cuts message delivery time between peers by 50 %.

P2P technology can put control of important tasks into the hands of key users, tie groups of users together in new and flexible ways, and ultimately, leverage your networking and desktop systems investments dramatically.

In fact, many see P2P technology driving the way we compute, develop, host, acquire and manage business applications in future.

The Fundas

Peer-to-Peer networking can be defined as “a networking architecture that facilitates direct person-to-person communication, collaboration and information sharing”. Intel currently defines P2P as “sharing of computer resources and services by direct exchange. When talking about corporate applications, it’s critical to focus on exchange between individual users rather than the devices and protocols — or else risk missing current benefits P2P can offer companies. Without a focus on users, corporate users might reasonably see P2P technologies as impressive but useless when it comes to actual business.

In practical terms, that means analyzing technologies by what users can do with them rather than how the work gets done under the covers. Next, as we are using the term, a peer to peer application is any application which relies primarily on a serverless architecture or tends to reduce the number of hops between end users dramatically. In other words, the word “peer” implies both two computing devices and two or more individual persons involved in an exchange. 

Advantage: P2pPeer-to-Peer (P2P) is a network architecture in which each computer has equivalent capabilities and responsibilities. This differs from client/server architectures, in which some computers are dedicated to serving the others. P2P enables the direct exchange of resources and services between computers by eliminating the server. Faster 
Delivery: P2P eliminates one leg of the communication highway (see diagram), thus at least doubling the speed of interaction. P2P also eliminates the need to temporarily store information in databases to relay information, which improves performance 5X to 10X. Greater 
Bandwidth: Under Client-Server, the Server is the bottleneck. The Server cannot handle everything the Clients can throw at it. As a result, Client-Server systems must drastically curtail communication and service quality. P2P bypasses this limitation by eliminating the Server, allowing the Peers to transmit 100X the amount of data.  
Infinite Scalability: In P2P systems, each desktop Peer brings all of the computing resources required for it to interact with the network. As a result, a P2P network can grow from 100 to 10,000 people without maintenance and scalability issues. 
Cost Savings: By eliminating the server, database usage and scalability issues, P2P systems are less expensive to install and maintain over time.

What P2P Can Do

As with any new development trend, P2P development borrows from a broad range of protocols and technologies which vary, depending on the specific objectives of the developer. 

To organize things usefully, we’re breaking down corporate P2P apps into four functional categories: managing and sharing information, collaboration, networking/infrastructure and e-commerce enablement. We believe that this may make it easier for IT buyers to consider what problems they hope to solve. 

Managing and Sharing Information

P2P technology offers the promise of new information management capabilities. In many cases, it does so by enlarging on technology models that already exist. Given Napster’s role in kicking off the P2P industry, it’s hardly surprising that this is one of the most active categories in the emerging P2P space. File-sharing products offer one of the simplest paths to adoption, as many of these technologies require little more than a download and a few minutes of configuration to use.Among other things, P2P products coming to the market can search a network’s worth of linked PCs, serving as the next logical step up from existing search engine technology.

Enterprise search engines are very valuable in their own right, but they only search servers. P2P solutions can allow transparent access to PC-based documents, at least those placed in public areas by end users, removing a layer of control and point of failure typical of traditional file-serving mechanisms.

The P2P interface from BadBlue, for example, turns PCs into search-enabled web servers. Using the BadBlue client, a tiny web server, users can configure shared folders and files and share spreadsheets in real time without having to download. Another model comes from Endeavors Technology, whose Magi Express is an open source thin server, based on the Apache web server engine, allowing common users to share, store and retrieve documents from common applications such as Microsoft Office. Other file-sharing players include Mangosoft, OnShare, OnSystems and a host of open source efforts.

Roku Technologies, for its part, offers a platform offering not only collaborative file sharing but also remote access to and control of files. Roku products — developed with partners — use P2P to remotely control a PC desktop using any web device. 

In still another twist on distributed information management, 100x has extended the P2P concept to database functions, developing what it says is the world’s first enterprise-class distributed relational database management system. For just about any of these products — as well as the collaborative technologies mentioned in our next section — secure transfer of information will obviously be an issue. One might argue that researchers consider it to be the key issue. Microsoft Research’s systems and networking group, for example, is at work developing its FARSITE P2P technology. Farsite is a serverless, distributed file system that does not assume mutual trust among the client computers on which it runs. Logically, the system functions as a central file server, but physically, there is no central server machine. Instead, a group of desktop client computers collaboratively establish a virtual file server that can be accessed by any of the clients.

Collaboration

Many companies have a love-hate relationship with groupware/collaboration platforms. Most love the productivity-boosting effect of a successful installation, but may find that usage drops off over time as people switch back to familiar (and interestingly, person-to-person-based) e-mail to get work done. What’s more, in the age of the company without walls, traditional groupware is less impressive as it may not be the tool of choice to communicate with partners, customers and suppliers. The next generation of P2P solutions hopes to close that gap, integrating various forms of person to person, serverless collaboration tools offering instant messaging capabilities to file sharing and project managers.

P2P groupware developers are also boasting secure methods to transport this information from one PC to another, across both traditional networks and the Internet. Among the best known solutions in this general category is the collaboration tool developed by Groove Networks, a company founded by Lotus Notes developer Ray Ozzie. Groove is designed to allow individuals to communicate directly and allow businesses the ability to communicate securely online with key customers and partners.

Perhaps the most complex products available in this space are aimed at the big-dollar network integration market. If distributed P2P app development calls for new tools, these companies are prepared to provide them.

Products like Cytaq’s enterprise software, for example, use P2P technology to help businesses connect all of their disparate technologies, platforms and computing devices at the edge into a connected whole. Another example is the networking infrastructure software developed by Proksim which, like Cytaq’s, is designed to extend apps from isolated servers and PCs out across every device touching the network.

And more is on the way. For example, OpenDesign Inc, a venture backed by former Microsoft CTO Nathan Myhrvold, will reportedly start testing a development platform this quarter which it says will let companies quickly write peer-to-peer apps based on reusable components.

Another subset of P2P vendors is hoping to evolve new, more efficient methods of network-based information sharing. For example, P2P vendor WareOnEarth’s Hypership technology is designed to offer secure delivery of data across untrusted networks.

Enabling Commerce

Still another rapidly-evolving category of P2P applications falls under the e-commerce umbrella. These apps, designed to enable e-commerce in new ways, stand to turn existing web commerce models on their heads According to an estimate by AMR Research, B2B commerce will generate $5.7 trillion in revenues by the year 2004. Not surprisingly, this has drawn scores of website builders and software companies into the net market maker arena, along with related financial ventures.

To date, many business models have stumbled, dogged in part by problems in coordinating the activity of buyers and sellers. But P2P technology, as it turns out, may provide answers to some of the key problems Net markets and other B2B e-commerce ventures face.

We’re talking not only about technical problems in software development, but also process problems which have until now undercut the profitability and even the viability of B2B e-commerce and exchanges. For simplicity’s sake, we’re also including supply chain coordination-type applications in this category, though they are undoubtedly deserving of their own analysis.

At present, some of the more visible emerging P2P commerce applications are aimed at simplifying financial market transactions. But smart enterprise managers will need to understand how trusted B2B payment mechanisms evolve, and may find short-term uses for these technologies in their current business practices.

Software developers have begun rolling out products that re-cast the net marketplace/exchange as a set of P2P transactions, rather than an online “community” mediated by servers and human guides. 
Companies such as Biz2Peer say that P2P architecture can not only facilitate buying and selling, but also streamline the supply chain. Looked at in another way, P2P companies in this space expect to offer the benefits of an e-commerce engine, an extranet and to some degree, a self-service customer care package rolled into one.

Other technologies worth watching, though more directly of interest to consumer transactional sites, are technologies from firms like Verai, which have set themselves up as mediators of a different P2P, “person to person” commerce. 

Verai, for its part, recently announced the release of a free wizard allowing any website to provide its visitors with a P2P credit card payment service. Such direct person-to-person payments may soon be the medium for small corporate transactions, skipping over the departmental credit card and accounting codes in one leap. A reach for the bureaucracies involved, perhaps, but not a leap of common sense. Be prepared for this approach to explode.

The Future

In all likelihood, P2P may be a buzzword which will lose its current flavour in the near future. However, the P2P mindset, which is a logical and evolutionary development for networking generally, isn’t likely to go away.

There could be a number of problems, mostly legal issues, which could hamper the P2P movement, adoption of technology, advent of new products. Nevertheless, it is clear that the technology is here to stay.

And in the coming months, many more companies including IBM, Microsoft and Intel will unveil their own P2P solutions, their own architectures and even products.

What is most important for any technology to click, especially in the enterprise business space, is the backing of at least one strong enterprise solution provider. P2P conceptually has many friends in the enterprise space. This could mean a bright future for it in the coming days.





Added on July 7, 2001 Comment

Comments

Post a comment

Your name:

Comment: