Business Strategy

Having looked at the fundamentals of peer-to-peer computing and the emerging landscape in terms of vendors and standards, we can now focus our attention on how to develop a business strategy around peer-to-peer with knowledge of various industry scenarios to which it can be applied and its current benefits and challenges. The primary benefit of peer-to-peer is the improved utilization of existing assets, which is something that should be of significant interest to businesses in today's economy. Companies can reduce costs by using existing assets such as desktop computers more effectively, thereby not having to purchase new hardware to support new initiatives. However, the main challenge for peer-to-peer is the security and management of applications. Without the centralized control of client/server architectures, information and transactions are less well-protected and can move around more freely on the network between peer computers. This makes the information more susceptible to loss or compromise, thus raising the issue of security.

Industry Scenarios

The industry scenarios for peer services break out into peer-to-peer applications for distributed computing to handle large computational tasks, and collaborative applications to facilitate ad hoc workgroups. Some of the business processes that can be enhanced and extended through the use of peer-to-peer collaborative platforms include purchasing, inventory control, distribution, exchanges and auctions, channel and partner relationship management, and customer care and support. Use of distributed computing cycles can be applied to any large computational tasks across a large number of industries. The following scenarios should help to illustrate some of the many applications of the technology to reduce costs and improve productivity.

Distributing Computing for Engineering Design (Intel "NetBatch")

As early as 1990, Intel deployed a fault-tolerant peer-to-peer distributed middleware capability called NetBatch for its Engineering Design Applications (EDA) environment. NetBatch has enabled Intel engineers to become more productive by using the collective computing power of workstations and systems across the Intel environment both in a local area network and wide area network scenario. This has enabled Intel to gain higher throughput and shorter run-times for their computing jobs and to run more complex jobs. The company estimates the NetBatch initiative has saved it hundreds of millions of dollars. In one instance, NetBatch helped saved millions of dollars in capital cost, by allowing engineers to tap into this global processing capability as part of one of their Pentium Processor (P6X) chip development projects. Without NetBatch, Intel would have had to purchase twice the local computing resources for the project. Over a five-year period, Intel has also noticed that workstation utilization at certain design sites has been sustained at the 70 percent level as opposed to the 30- to 40-percent level prior to the deployment of NetBatch. This has helped to optimize the utilization of existing resources before requiring new investments in further processing capability. It has enabled Intel to achieve more with their existing resources a compelling story for the adoption of peer-to-peer computing for businesses that have compute-intensive tasks as part of their product design lifecycle.

Peer-to-Peer Knowledge Management for M&A Activity (Baker and McKenzie)

The law firm Baker and McKenzie used peer-to-peer knowledge management software from a company called NextPage in order to improve productivity for its Mergers & Acquisitions practice. The company is a global law firm with 62 offices in 35 jurisdictions and over 3,000 attorneys. The NextPage Matrix product was used to connect attorneys distributed around the world with the people and information they required in order to perform complex merger and acquisition transactions. Examples of content made accessible by the system includes legal precedents, local regulatory issues, and industry-specific research. The benefit over traditional content management systems is that the application makes content available no matter where it resides within Baker and McKenzie offices around the world. It also helps connect the firms' clients to the business process as they join the collaboration. Additional business benefits have included reduced requirements for travel before M&A deals are signed and the ability to train new associates on transactions via the collaborative workspace.

Supply Chain Collaboration

Within the supply chain, peer-to-peer collaborative tools can be applied as an extension to existing enterprise resource planning and supply chain management systems. They can be applied for interenterprise ad hoc collaborations related to exception-handling or other people-oriented tasks that are not well-accommodated by process-centric packaged applications. They help to provide a platform for groups within the supply chain across various companies to converge rapidly to discuss and resolve issues and to store a permanent record of the interactions for future knowledge management. External information that impacts a supply chain process but which is not contained within the core-packaged supply chain management systems can quickly be disseminated across supply chain participants with file sharing, cobrowsing of Web sites, and instant messaging.

Partner Relationship Management

Since partner relationship management is highly collaborative and ongoing, it is an excellent candidate for a collaborative platform either based on a client/server or peer-to-peer architecture. Peer-to-peer can help by empowering all participants to make information available from their own machines and to configure the peer groups and collaborative tools they want to utilize. Costs can also be saved in this model due to the reduced requirements for any server infrastructure to be established along with the IT administration requirements. For small group collaborations among many different partners, the peer-to-peer model can be highly efficient. For larger, more institutionalized relationships and strategic alliances, the model should usually be combined with client/server models such as extranet Web sites for greater control over the formal relationship and information exchange.

Benefits

The business benefits of peer services fall into two main categories: collaborative benefits and computing benefits. The collaborative benefits are mostly related to the business benefits around knowledge management. Applications that are developed using the peer-to-peer computing model can enable knowledge workers to be more productive in a variety of industry and functional areas. Products such as Groove from Groove Networks provide knowledge workers with the ability to create and manage ad hoc groups to conduct business processes in a far richer, collaborative environment than the disjointed phone, email, and fax interactions that are typical of today's office environment. As Groove Networks states, the technology helps people in different companies or different divisions reduce their "cost of connection" in collaborating with one another and performing their work.

The second business benefit revolves around the exploitation and optimization of existing computing resources. As illustrated in the Intel NetBatch example, businesses can generate large cost savings by maximizing the use of their existing infrastructure. Workstations across the organization can be tapped for their unused processing cycles in order to perform intensive processing jobs as part of product design activities, drug research activities, or other intensive tasks such as seismic data processing in the oil industry. The processing power that can be realized from these large connections of peer computers on the network, all applied to the same task, can be far greater than even the most powerful mainframes. This is especially true when the network is opened up to outside participants, and thus a larger pool of potential participants, as in the Intel Philanthropic Peer-to-Peer program and its programs for drug research.

Challenges

Some of the challenges to the peer-to-peer computing paradigm are similar to the challenges faced by other infrastructure technologies. It needs to gain acceptance by minimizing technology risk and maximizing business benefits. To minimize technology risk, several topics need to be addressed. The main items are security, management, and interoperability. For business audiences to feel safe about using solutions built upon peer-to-peer architectures, they need to have the same level of confidence as they do with their current IT infrastructure and applications in terms of ease of use, security, reliability, availability, performance, manageability, and maintainability. Business users will expect to retain full control over their desktops and be able to fully configure which resources are made available for others on the network and which users are granted permission to access those resources.

Since the challenge is really an information technology management issue, business end users may be unaware of the potential challenges to deploying peer-to-peer computing. As an end user application, peer-to-peer architectures will seem more user-friendly and convenient than most server-centric applications since the administration responsibility is shifted to the business users themselves. It is information technology managers who will have the greatest challenge in supporting these applications. With no central control, or at least with limited central control, they may well be reluctant to allow these types of applications onto their networks. Software companies in this space will need to carefully educate their audiences as to the security and manageability of their products in order to penetrate the market.

Strategy Considerations

Given the strengths and weaknesses of peer-to-peer computing, the initial strategy focus should be two-fold. First, one should observe where to gain the most benefits from highly collaborative processes. Second, one should look for computing tasks that are highly intensive and which can be distributed in a parallel processing environment. For most businesses, the first area is most likely to be the area of opportunity. Many businesses have collaborative processes that can be enriched with ad hoc peer-to-peer platforms. This can help reduce IT costs and resource costs in terms of IT administration. In an era of focus on cost takeout and performance improvement, this type of solution can enable business users to become more productive without adding costs into IT departments.

The second area of potential benefit for the business is via distributed processing for complex computing tasks. This type of implementation will require a much higher level of upfront planning in order to both identify computing tasks that are suitable candidates and to plan the design of the solution. While collaborative platforms can be purchased off the shelf and are often Web-based and thus easy to install and configure the design and implementation of a distributed processing solution for complex computing tasks can be lengthy. The business will need to determine which computing tasks to distribute, which and how many machines to distribute the tasks across, and whether or not the tasks can indeed be run in parallel by a number of machines simultaneously. Even if all these factors are determined, the solution will still require the careful design of a controlling application to send and receive computing tasks among the peer computers involved in the solution and to maintain auditing and reporting information. Businesses embarking on these types of initiatives should look for standards-based toolsets such as those from the Peer-to-Peer Trusted Library or from Sun's Project Juxtapose.

In summary, while the business benefits may be apparent, it is important that the business looks for packaged solutions and standards-based frameworks to target these opportunities. This is far preferable to tackling complex development tasks from the ground upwards with standards that are still evolving.

It may also be possible and beneficial to incorporate peer-to-peer considerations as part of Web services initiatives. Your business may well be engaged in building components of business functionality to be exposed as Web services on the network. If this is the case, then peer-to-peer may enable those services to be available not only to enterprise servers but also to personal computers, devices, and appliances at the edge of the network. In this manner, peer-to-peer computing can help to make Web services device independent and to maximize the usage of all computers on the network and not just server-based machines.

Estimating Results

As with any other implementation, return on investment for peer-to-peer computing is application-specific, so we will break this discussion into two discrete scenarios: a scenario for collaboration and a scenario for distributed processing.

In the first scenario, return on investment for peer-to-peer collaboration can be determined by measuring the cost reduction when compared to deploying a similar application using server-centric techniques, or when compared to more manual, disjointed collaborative techniques such as phone calls, emails, and faxes. For example, imagine a collaborative hub that includes instant messaging, text-based chat, document sharing, and threaded discussion that is rolled out to 50 employees as part of a six-week internal team project. The return on investment can be determined by the increased productivity of the employees, plus the cost savings from reduced administrative support, divided by the cost of the peer-to-peer platform.

If we assume a $100 per year end-user license for the peer-to-peer desktop software, then the total software cost for the 50 employees in our example is $5,000 per year. We might also assume that, in this case, the administrative support required is reduced from one day to zero due to the self-administration by the participants. Administrative support in the alternate server-centric scenario might include the time spent in setting up user accounts and access privileges and in creating the required discussion forums and document-sharing capabilities. The cost savings for administration might be $1,000 for one day of an administrator's time. This can be discounted from our calculations, however, since the end users will need to pick up some of these administrative functions themselves. Finally, we might assume that end user productivity is improved by five percent, or two hours per week, due to the improved collaborative capabilities of the platform. For 50 end users working for six weeks, this amounts to a time savings of 600 hours throughout the course of the project and a cost savings of $60,000, assuming a $100-per-hour, fully loaded cost per employee. The return on investment in this example is 1,200 percent with the cost breakeven coming after the first few days. When dealing with knowledge workers, the return on investment can be considerable due to the value of their time. Even small productivity improvements can go to straight to the bottom line.

The general formula for return on investment for peer-to-peer collaboration in this example is as follows:

Return on investment (collaboration) = (Cost savings from end user productivity + Cost savings from reduced administration) / Cost of peer-to-peer platform software

In the second scenario, the return on investment for distributed processing is equal to the cost savings from reduced hardware purchases divided by the cost of the peer-to-peer project in terms of hardware, software, and systems integration resources.

Return on investment (distributed processing) = Cost savings from reduced IT purchases / Cost of peer-to-peer project

If the business can realize substantial cost savings in terms of new hardware investments by using existing assets, and the project can be planned and implemented quickly at minimal cost, then there is a good case for a return on investment. The technical feasibility of the project should be examined very carefully, however, since it often requires a complete redesign of how current processing tasks are conducted and raises several new issues for IT departments to deal with. The ability to make the investment last for multiple years is another factor that should be considered. In the Intel example, the NetBatch initiative has been in operation since 1990 and its longevity has ensured a strong return for the company.