NearPath is a sophisticated set of technologies that provides automatic site-to-site synchronization of customer uploaded files, best site selection - directing downloads to the site/server that is nearest to the user, and dynamic site-to-site monitoring and load balancing that ensures that downloads are directed to sites/servers that are always operational and never overloaded. The technology provides a completely distributed server model with no single points of failure.
As a customer uploads a new file to any server in the DTI network, NearPath automatically and instantly copies the new file to each of the other sites/servers in the network. Typical synchronization time is just a few seconds or less. Site-to-site synchronization involves a number of different technologies which provide job queuing/scheduling, parallel processing with conflict resolution, and failure detection and retrying. In addition, any synchronization delays are mitigated by another algorithm which can redirect downloads back to the original upload server - ensuring that files that have just been uploaded are immediately available for download across our entire network.
BEST SITE SELECTION
For each download, NearPath uses a proprietary and highly accurate set of databases to translate the user's IP address into a set of latitude and longitude coordinates. Using the "Great Circle" mathematical formula, these coordinates are then combined with the latitude and longitude coordinates of each site/server in the DTI network to create a set of distance vectors (the distance between the user and each of our facilities). These distance vectors are sorted in nearest to furthest order and the user is then automatically directed to the nearest operational server.
SITE-TO-SITE MONITORING AND LOAD BALANCING
Servers in the DTI network are organized into server groups. A server group is defined as a group of servers that share identical content. There are one or more servers for a given server group at every location in the DTI network. Each server monitors both the reachability and load of every other server within the same server group. This monitoring is performed every few seconds and a database is updated with any changes that are detected. This database is then used along with the distance calculations mentioned above to find the nearest operational server.
NearPath technology is applied at multiple levels, starting with the DNS (Domain Name System) name resolution. At the DNS level, the user is directed to what we refer to as the "first contact" server. Since the machine that is performing the DNS lookup is not necessarily the same machine the user is on and not necessarily physically close to the user's machine, NearPath at the DNS level is considered 'a really good approximation', but not the final answer. The final answer comes when the user contacts the server at the HTTP protocol level. At this level, NearPath looks at the size and nature of the file being requested along with the distance to other servers in the network and then decides whether or not it may be higher performance to issue a redirect to a closer server, or to handle the HTTP request immediately. It is usually less optimal to redirect a small file request than it is to just handle it directly. Larger files may, however, benefit from being served from a closer server. In most cases, NearPath at the DNS level has already directed the user to the nearest server, so redirects at the HTTP level are rare.
OTHER USES OF NEARPATH TECHNOLOGY
The NearPath IP to user location techology is additionally used to generate by-country and by-state download demographics graphs which are made available to our clients via the Customer Portal website. These graphs show where your users are located and can be of great value in establishing and evaluating the effectiveness of targeted marketing efforts.
Copyright 2004, Download Technologies, Inc. All Rights Reserved.