by Vincent S
Today I am posting about an article I found in the ACM Digital Library. The article was a peer reviewed journal submission presenting a solution on how to manage a database in a mobile ad hoc network. For those you are not familiar, an ad hoc network is a wireless network in which no network management device is present (switch, router, hub, bridge) and all host devices have wireless capabilities. The reason for such a network to exist would be for circumstances in which a wired network would be difficult to set-up and maintain. Such a scenario could include in battle or more relevant to this class would be for disaster recovery for a business in an emergency situation. Possible challenges that might present themselves to a database administrator are easily intercepted and comprimisable data, lack of data integrity, and network collisions that might cause the loss of data due to lack of network management. The authors of the article posted various solutions to these challenges.
First off, the architecture of the network would be of vital importance. Most wireless cards only have a secure range of 100 feet. The consequence of this is the fact that many hosts within the network would not be able to talk to each other directly. Hosts would have to communicate indirectly by sending packets in hops, meaning the info hops from one host to the next until it reaches its destination. A typical setup could include twelve hosts. Two hosts will act as serves and will hold the entire database and the metadata and would be placed at opposite ends of the topology. The topology would consist of a “RING” and would have an additional nine to ten hosts for which data can be processed. The host themselves would only contain partial images of the network needed for whatever aspect of the database that the host is responsible for maintaining.
In terms of updating the information and ensuring integrity of information and security, a VPN (Virtual Private Network) is used to create a “TUNNEL” for information packets to travel in-between. Each packet is assigned a randomly generated key that is encoded with an algorithm for the sake of encryption. This makes sure that data was received from a legitimate source and has not been tampered with. Each update to the database is treated as an individual transaction and is sent to the server hosts to be qued. Each update must be individually verified and added to the main database.
As you can see, there are multiple concepts covered in the class. This is important because it shows the furthering capabilities of database management. In the world of the computer networking industry, ad hoc networks are common for minor business applications. However, they are not too commonly used for the management of sensitive information. The fact that it is being used for database management and that they have found possible solutions to make it safe and possible is quite fascinating. Two such consequences of this setup are the ability for database administrators to telecommute, and the increased capabilities of a business to create a disaster recovery plan. With this setup, a company can create a temporary network that has copies of database images from the main server. If a main server goes down, the ad hoc network will contain the images of the updated information for the day and can be used to recover data. The integrity of this network would not be affected as it does not rely on a wired connection.
In closing, as stated in my reflection, this concept offers an array of possibilities. Although database management is not my field of choice for computer information systems, it obvious that is becoming a high demand lucrative field. It would be even to my benefit to learn about managing a database in an ad hoc environment and more about the planning and science involved in disaster recovery.
Gruenwald Le, Banik Shankar, Lau Chuo (2007). Managing real-time database transactions in mobile ad-hoc networks. Retrieved from http://www.springerlink.com/content/5k4788j051g52542/fulltext.pdf