Main Server Diagram
The design is a client server application approach. The system has two layers of applications, GUI and API that connect to a server program.
Server
The Global Datastore consists of the knowledge base and reference data.
The data in the knowledge base can be edited (after training and with knowledge and caution)
by using the Datastore Editor.
The reference data is primarily made up of address data, sourced from postal authorities,
mapping data providers, and other in-country address data sources.
We have previously looked at the UPU (Universal Postal Union) data,
but although the coverage is fairly good the updates are pretty infrequent.
Having the direct relationship ourselves allows us to be proactive in constantly looking
for newer and better sources of data.
The second section of the reference data covers enhancement datasets.
These are datasets that we can provide which will add extra information such
as latitude/longitude, business information, and geodemographic information.
Finally through the use of the Data Compiler we also have the ability to add
user datasets to the Datastore, allowing customer-specific information to be
used during the validation and enhancement processes.
All reference datasets are read-only. This is to ensure t
hat we maintain the integrity of the supplied data, and it's
a stipulation from a number of our suppliers (USPS, RoyalMail, etc)
that we keep the data secure and uneditable. However, what we do provide is the ability for a customer to specify exception data, which may include specific new-build addresses that haven't made it into the reference data yet. Obviously this needs to be used with caution - if this is desirable for a customer any data added must be thoroughly checked for validity. By keeping this data separate it also means that updates are easier - replacing the underlying reference dataset from, say, the USPS, doesn't overwrite the exceptions that have already been entered
Cluster
The clustering functionality allows call-routing with one of three different loading techniques: redundancy, load optimizing and load sharing.
Clustering capabilities allow for peaks and troughs in server usage
With the clustering capability, servers need only be installed with the locally used country reference data, and route calls for other countries to other servers, including an OnDemand server.
Network
Connection from a client level to the server can be:
- Internal to one PC
- Over a local network, thus allowing multiple clients to connect to one server
- Over the Internet to either a hosted OnDemand server or a customer server stored at a different location.
API
The first layer of the client approach offered. All APIs apart from the Web Service approaches offer the same functionality and can be used to develop bespoke address processing applications. APIs are always required to be stored on local PC even when calling a server not running on the local PC.
The Web Service approach offers the core functionality and can be accessed via a WSDL interface. This approach allows address processing without installing files locally.
The Java Script RPC is a simple means to integrate a rapid address generator with minimal coding. This approach downloads script to the customer website which works seamlessly in their forms.
Application
Applications developed that use the functions available through the APIs. All functionality offered in these approaches can be replicated in bespoke applications.
Interactive: Is a rapid key, address entry tool that generates accurate postal addresses from as few as eight keystrokes. It integrates seamlessly and rapidly into contact centre and other software environments, and can be used on your website for the fast and authentic capture of visitor addresses
Batch: is an advanced, automated data processing application used to cleanse and enhance global name and address data. It can function either on the desktop or within a specialist data processing environment. This intelligent software standardises, validates and corrects mixed data from up to 237 territories simultaneously. It can enhance name and address data with additional data sets such as geo-positioning and time zone data.
INSTANTIntegration: enables the seamless integration of Interactive into third-party applications in minutes. Using the simple drag and drop set-up tool or key sequence recorder, Interactive can be integrated to send accurate data instantly to other applications, removing the requirement for lengthy development consultancy
|
