Surety is performing system maintenance this weekend. Electronic date stamps on new Prior Art Database disclosures may be delayed.
Browse Prior Art Database

Determination of the Geographic Server Location

IP.com Disclosure Number: IPCOM000238993D
Publication Date: 2014-Sep-30
Document File: 5 page(s) / 101K

Publishing Venue

The IP.com Prior Art Database


Virtualization enables the separation of image layer from the underlying hardware platform. Modern hosting providers can even move images between their data centers for load balancing reasons especially in Cloud Computing. The customer might be interested in the geographical location of virtual sever images and data for several reasons: - Moving an image to another location can change the application response times due to network latency changes. - Legal or company regulations may require to store data in certain country or political units. - For disaster recovery purposes a set of servers or certain data copies must not be located in the same data center or even same country. - Audits may require information about the image location for the reasons above. The requirements above can be translated into policies, containing conditions and scope of servers. For this reasons an automated process (supported by a tool) validating the policies on the servers would be useful.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 29% of the total text.

Page 01 of 5

Determination of the Geographic Server Location

The principle idea is to use network triangulation to determine the geographical location of a virtual server. The location of a point can be uniquely determined if the distances to (at least) three reference points are known. The location is given by the common intersection point of the circles.

The principle idea is to locate a virtual server using thenetwork latency to reference servers. The latency is obtained by network utility ping and corresponds to the length of the network path.

Certain considerations are required to achieve a meaningful result:

1. The measurement of the latency (as a scale for the distance) is not exact.

2. Latency cannot be translated linearly into distance. A functional relationship between latency and distances must be established based on measurements preformed on reference servers.

3. To reduce the error margin more than three reference points should be considered. Interim results are obtained from triple of reference points. Interim results can be compared, data out of range can be dismissed, and consistent interim results can be combined in a weighted average to yield
a more precise final result.

All the necessary steps can be fully automated in a tool. The tool will present a map showing the geographic location of the virtual server hosted in the provider's data centers.

Figure 1: Architecture Overview Diagram

Agent - Latency data collector

The agent performing latency measurements using ping and/or traceroute is running at least on the reference servers which are in control of the customer's organization. By providing the distance between the reference servers in terms of latency, the scale to convert latency in ms to distance in km is set.

Agent Configurator

The agent configurator configures the agents with the IP addresses of the references servers and the latency measurement time schedule.

Data Processor

The data processor maintains a database with current and historic information. The data processor receives the latency measurement results and computes the geographic location of each virtual server


Page 02 of 5

and stores it in the database "Triangulation Results" (TRDB).


The console provides the current and a historic view on the locations of the virtual servers.


The reporting component draws reports from the information of both databases.

Functional Relationship between Latency and Distance

By pinging each other the reference servers provide information about the distance in the network in terms of latency. By combining this information with the known information about the geographical distance a correlation between latency and distance can be established.

Polynomial Regression
In the example below the eleven data points are lead to the following relationship between latency L and distance D using 2nd degree polynomial regression: L = 0.3 D2 + 30 D - 300

Figure 2: Using a polynomial regression to conver...