Enough buzzwords to get by

As we discuss how Jtrix implements its solution we will need some buzzwords to help us, but will keep them to a bare minimum. Meanwhile you Table  provides a key to our diagrams.

Table: Images used in this document. The terms are defined in the text.

Node	A node and the servers it spans	A netlet
A netlet bound to a node	Two netlets bound together	A service
A descriptor	A netlet descriptor and its netlet	A warrant

Service

Any on-line activity performed for others and which is generally available is considered a Jtrix service. By ``generally available'' we mean that the business in general is the service, not a single transaction or a series of transactions taking place under a single agreement.

A service includes activities such as Harry's Web-based mail or chat provider, credit card authorisation or news headlines. But a service also includes access to resources such as disk space, CPU time or bandwidth.

Node

Any Jtrix application runs in a Jtrix environment which manages various services. We can imagine that this environment is based in a single server or PC. But this need not be the case--it could physically exist over several machines working together. For this reason we do not refer to a Jtrix environment as a server or a PC. We call it a node. A node is one or more machines working together to provide a single environment for applications. It may be run by an individual, a department or an ISP. Regardless of whether a node is one machine or one thousand the application does not need to worry about it.

Netlet

A Jtrix application is made up of one or more program components. There will often be a single core component, with each service it buys into being fulfilled by another component which attaches itself to the original one. Each such component is called a netlet. A netlet is a program or program-fragment. Any one netlet may use any number of other netlets to fulfill its task, or it may be stand-alone.

It is important to understand these concepts now. To summarise: a netlet is a program component, possibly a large application, which lives on a node. A netlet will need services from elsewhere, so it will attach other netlets to itself to use these services.

We therefore also get the idea of a service netlet. This is simply a netlet which is used to fulfill a service, which attaches itself to the first netlet which wants the service.

The concept of services being generally available is very significant for Harry and his ASPs. For example, Zippy Filestores may provide secure on-line data storage for applications with their ZFData system. Harry can buy space up to 200GB on the ZFData system, and access it periodically over a six month period from various locations around the world. Someone else may buy 10MB for one week and access it continually for that time from two locations simultaneously. Although it is being used by several people and system, only one service is in operation here: the ZFData service.

We will also come across some other buzzwords, although we will define them later: service binding (Section ), descriptors (Section ), warrants (Section ) and wallets (Section ).