Peer-to-Peer in Insurance? Not Yet

Q: Could insurers benefit from peer-to-peer (P2P) computing? Will there be wide-scale adoption of P2P in insurance?

A: Max Drucker, chief information officer, eCoverage (San Francisco): Unfortunately, for the insurance industry, I don'tbelieve that P2P is going to make a major impact. I don't buy the argument that P2P can provide a more efficient distribution architecture, either. If you are looking to serve up content to a large audience and do not want to make the capital expenditure...simply use Akamai or the other caching servers that are far more efficient than any P2P network could ever be.

A: Howard Sachar, manager, IBM Global Insurance Research Center (Hawthorne, NY): There are several benefits from P2P computing. Since the resources of a carrier's PCs are under-utilized for a majority of the day-after business hours, for instance-there is computing capacity available that could be used in a dynamic on-demand fashion.

For internal operations, P2P allows independent software agents to communicate directly. Working in parallel, a distributed network of intelligent software agents can service complex objectives. For instance, a P2P network can work to search for insurance products that meet the requirements of a customer. Through these types of application scenarios, P2P can potentially reduce the need for additional centralized computing resources.

The P2P model can also increase collaboration within the extended insurance sales and processing chains. For example, the claims process is a multi-party collaborative activity including customer, agent, adjuster, and others. Com-munication among these parties can be improved with P2P. For finalizing sales, insurers will most likely not be early adopters of P2P. Instead, P2P will have early commercial impact on areas where commodity items are exchanged. Insurance products usually require some customization. However, we see now and will increasingly see P2P used in matchmaking-finding the right product or service for customers.

A: Kimberly Harris, senior research analyst, insurance, Gartner Group (Stamford, CT): While P2P computing is receiving attention in other industries, the value it may add to the insurance industry is limited. The goal of P2P computing is to share information or collaborate between individuals in an informal manner. It is simply a modern extension of the news groups and message boards that emerged in the 1980s.

A: Thierry Gabadou, global business development manager, Insurance Sector, Sun Microsystems (Palo Alto, CA): P2P computing can be thought of as a natural extension of network computing and the Internet. In the same way that the insurance industry reaps the benefits of network computing today, we anticipate future opportunities and developments in P2P computing as the technology matures and emerges further.

As it exists today, P2P leverages economies of scale for highly intensive CPU resources, though the insurance company typically requires efficiencies in high-volume transactional processing, where each is non-complex. Broad adoption of P2P will be determined by the ability to call on any device, (server, PDA, PC or smart card) on the network as a peer.

Q: What are the potential benefits of P2P for insurance?

A: Sachar: P2P offers the potential to reduce costs by getting better utilization from lower-priced capacity, and gaining performance from direct communications between parties. It offers a model for simpler modes of collaboration and the sharing of both structured and unstructured data (e.g., voice, image, and free text). There are currently offerings that attempt to utilize unused computing power for financial computations. However, for the potential to be fulfilled it requires that calculation and storage be productively discovered and utilized.

Standards will also need to evolve before P2P can reach its full potential in insurance. PCs need to all be talking the same language or have reasonable abilities to translate. Insurance policies today are substantially more complex than, for instance, music files.

A: Gabadou: For the industry as a whole, the advantages of P2P network-based computing are linked to a highly scaleable infrastructure where data is shared between resources as the volume of business transactions increase. Computing and storage power has to grow with volume. If an insurer sells more policies or processes more claims than expected at any given time, it's much easier to turn the power up a notch on a few servers than ramp up and maintain 1,000 PCs.

Q: What technology is needed to make P2P computing a reality for an insurer? How is the technology different from what insurers already have?

A: Sachar: Improved technology for searching, security and privacy are needed for all P2P scenarios. Improvements in distributed workflow and e-markets technologies expand its potential use-the ability to create, manage and run distributed directories must keep pace. In an environment partially or fully dependent on the P2P model, separate businesses that support insurers will be built around certifying participants and running directories.

Of greatest importance is the industry's need to evolve to match the capability of the technology. The technology is decentralized, which implies independent action by equal participants. This is significantly different from centralized data repositories with controlled access.

A: Gabadou: From the standpoint of P2P network-based computing, a scaleable, secure architecture based on open standards is essential. The architecture will provide several benefits, including:

- Allowing for product developments in a timely fashion.

- Embracing anything with a digital heartbeat as a peer.

- Cutting administrative costs.

- Empowering channels to best serve their customers.

Q: What are some of the security concerns with P2P computing?

A: Drucker: The only real concern that P2P brings up should be resolved at a firewall level. Incoming traffic should be blocked at the firewall, so even if a user on the network runs P2P applications, security weaknesses that P2P applications might have should not really be an issue. The bigger problem can be the downstream bandwidth; as many users download files, the connection for other users can clog, as universities have experienced with Napster.

A: Gabadou: Security is of paramount importance in P2P computing-just as it is in any other form of network computing. Java used in this type of network computing was created with security in mind. Network computing has several advantages over traditional fat-client computing:

- A PC's operating system is not designed to handle mission-critical data.

- A PC is not as reliable as a server. If a PC goes down, data may not be available.

- One server is easier to administrate than a few thousand PCs.

CIOs and their teams should look for the open standards-based architecture. Such a model leverages reusable components that decrease time to market, and powers applications written once, to be used securely, over and over again.

A: Sachar: Implicit in P2P is the ability to directly communicate and share capabilities across machines. Giving access to files increases vulnerability to attacks (e.g., viruses) and altering of information. As P2P computing creates an anonymous distributed network of collaborators there is a greater need to authenticate participants.

The high levels of privacy and security needed in insurance will make most CIOs take a "wait-and-see" approach. They will need to securely fence the areas and functions that will participate in P2P environments and construct barriers for their core assets. CIOs must be prepared for significant organizational and cultural challenges as the environment becomes increasingly distributed.

A: Harris: P2P exchanges lack the efficiency obtained by using email, lack methods of accountability of who is distributing the information, and may not provide as secure methods of sending data and upholding privacy. Overall, P2P computing does not add any additional value to the insurance industry over current methods of computing.

Greg MacSweeney is editorial director of InformationWeek Financial Services, whose brands include Wall Street & Technology, Bank Systems & Technology, Advanced Trading, and Insurance & Technology. View Full Bio