Dit artikel bevat een functionale beschrijving van de Activator techniek. Deze tekst is alleen in het Engels beschikbaar.

Ontwikkelaars kunnen een PowerPoint presentatie downloaden, waar de technologie nog diepgaand in beschreven staat: Activator For Developpers

Onderstaand artikel is ook te downloaden als PDF-bestand: Activator Whitepaper

Anywhere, Anyone Networked!
The Activator: The way to Adaptive Computing

1. Introduction

We are on the brink of a new era in the evolution of Computer usage: Mobile and/or Distributed Computing as a structural part of the back-office. The ever increasing capabilities of Computer Hardware infrastructures has made it possible to build ever smaller general computing devices. Not only have these devices become smaller but also less expensive, bringing them in the financial reach of ever larger user populations.

Until recently, Mobile Computing was limited to special devices, often tailor made for the projects that these devices were intended for. Being so special, the mainstream IT department generally did not consider these devices as real computers or hardly being a part of the infrastructure: They were looked upon as foreign devices that needed little (data-entry) or no real interaction with the back-office platform. Mobile projects were often considered to be isolated.

Very recently, mobile devices have become more and more popular. More and more people are using these devices, either for their own pleasure or for business. This evolution caused IT departments to shift there attention more towards these devices. The increased power as well as the popularity has caused users to request more and better integration with the corporate data and information structures. Eventually these changes in attitude will lead to the acknowledgement that the mobile platforms, whatever they are used for, will be considered by the company as an integral part of the overall IT infrastructure.

Currently, mobile devices are called "mobile" because the user carries them around. When the user needs to connect the device with the back-office, the communication infrastructure used, regardless if there is a physical wire present or not, is rather inflexible. The layout of the network once the connection is established (= topology) is such that it looks like it uses wires. Indeed, most mobile devices are considered to be satellites of some user controlled 'host' (call it the users' desktop). Whatever physical network the device uses, it will for most (vital) services want to connect to that single host. As a consequence, most services on a mobile device are currently configured to be run on a fixed network layout or STAR-network.

The Internet has provided us with a different kind of network layout. On the Internet ANY device can connect from ANY location and request a communication session to ANY host machine in ANY location. The network is neutral with respect to the logical, application determined, communication that is being established and/or used. Also the calling device refers to the host that it wishes to connect to by specifying a logical network address. This logical address is translated by the network to the physical address of the host. The calling devices will never need to know that physical address. The network topology of the Internet is called a CLOUD-based network layout. There is no difference between the traditional client and server roles : callers and called are peers.

If one could take a snapshot of all communication sessions that are active at any moment in time on the Internet, one would see a rather chaotic layout of these communications. In fact, the Internet itself will (automatically) determine what the physical path of the data from the called (source) to the calling (destination) computer will be. Even stronger, the data will be chopped up in smaller parcels or PACKETS and the network could assign a different route for each of the packets. The network takes responsibility for the disassembly (in the source) and assembly (in the destination) of the transmitted packets. All this activity happens transparently to the application : the flow of data will appear identical on both sides of the communication.

Now that we know that the Internet brings a new view over networking, we can consider its impact on applications and application design. Since applications never use physical addresses of the hosts they wish to connect to, the network does not guarantee what host really actually acts as server. The assurance for finding the identity of the server, will have to be provided by other services on top of the physical network (e.g. Security services). At first glance one could say that it is difficult to imagine situations where the application does care to know the server it connects to. However these situations are in fact not so rare. E-mail for instance. If it were not because of Spamming, no user cares what E-mail server will pick up the message as long as it reaches its destination. Many applications really do benefit from this server-transparency. Techniques like load balancing (multiple parallel servers to serve a larger population of clients), roaming (a different server takes over the connection while the client moves through the network), auto-detection (the client broadcast on the network to detect what printers are available) are all based on network transparency.