Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Son of SAN - Storage Networking Technologies -- Ignore unavailable to you. Want to Upgrade?

To: Joe Wagner who wrote (2906)	3/10/2001 1:08:07 PM
From: Joe Wagner	Read Replies (1) \| Respond to of 4808

IP addresses have us driving in circles By Michael Tanne, XDegrees Inc., Special to ZDNet Saturday March 10 12:15 PM EST dailynews.yahoo.com I was late for a meeting in San Francisco the other day. As I was making my fourth pass around the block looking for a parking space to open up, it struck me that the way IP addresses are assigned to computers is becoming like parking in San Francisco. Much like tenants, servers are assigned permanent IP addresses. But with the tremendous growth in the number of connected PCs and devices, there simply aren't enough IP addresses to go around. So we have resorted to a system in which PCs temporarily use IP addresses borrowed from a pool. The problem is, the Internet's mechanism for finding machines, the Domain Name System, relies on domains assigned to fixed IP addresses. This leaves most PCs and other devices unnamed and unreachable. The Internet's venerable DNS has served us well these past 17 years, and Paul Mockapetris and the others who contributed to it deserve our thanks. But DNS is beginning to show its age. Many new applications, such as instant messaging and Napster now rely on connecting directly to PCs. If this approach is to succeed, we need another way to name and access PCs and other devices. The dark matter of the Internet Astronomers have concluded that the universe contains a tremendous amount of mass that we can't see. They call this "dark matter." In the same vein, Internet resources such as PCs, devices, files and applications without URLs or domain names are not visible and available to us. Clay Shirky, a Partner at the Accelerator Group, calls PCs the dark matter of the Internet. I'd go further to include files and Web Services. And DNS is powerless to illuminate them. As broadband lights up the edges of the network these valuable resources become more available, but they are still intermittently connected. This requires us to rethink our assumption that Internet resources are always available, and anything less is an error condition. Instead, we must accept that many resources, such as laptops and devices, are intermittently connected, and design our systems accordingly. When these machines go offline and return with different IP addresses, they ought to maintain a stable identity. If we can abstract resource names from their physical locations we can make them behave together as a more reliable, integrated network. Unfortunately, we don't yet have such an abstraction layer. For example, Napster presents users with long lists of identical file names that point to different copies of the same file. If the name were abstracted from the physical file, searching for "Angel of Harlem" would yield just one result for each distinct recording of the U2 song by that name. The software could then deliver the closest, fastest copy of the file to the user. What's in a name? What then, would an ideal resource name system look like? If it is to support network applications, a name system should have these qualities: An extensible namespace mechanism that is based on the URL and allows applications to leverage existing naming schemes. The ability to identify resources as they connect or reconnect to the network. The ability to identify all types of resources (files, services, people, devices and so on). An abstraction layer between logical names for and multiple physical copies of resources. The ability to verify that multiple copies of a resource are actually identical. An extensible resource name system, as outlined above, is key to the development of applications that leverage all available network resources. Network applications The Web and most enterprise computing rely on a client-server approach. Recently peer-to-peer technology has emerged as an alternative. In practice these two approaches will be integrated, allowing future applications to harness the strengths of each. Multidirectional communication between servers, clients, peers, devices and any other resources on the network would then become the norm. As network applications mature, they will provide new efficiencies and a broader set of network capabilities to existing and new applications. Innovative companies will deliver applications such as distributed virus protection, backup, and information sharing. Looking forward A name system and abstraction like the one I've described will allow us to access and utilize resources throughout the network more effectively--essentially illuminating the dark matter of the Internet. Next time we will discuss the similarities between quantum entanglement and dot-com financing. Michael Tanne is the founder and CEO of XDegrees Inc., a Mountain View, Calif.-based developer of secure, flexible infrastructure services for developers of enterprise and consumer Network Applications. For more information on XDegrees, please visit www.xdegrees.com.