Software Architecture
Nimbus is a Multi-tiered Client Server product which uses
Microsoft SQL Server 2000 or MSDE,
(a free product for small users), as its database. We use Microsoft Visual basic
for the applications and middle tier servers.
We chose Microsoft SQL Server due to the benefits
it could offer our clients. These include:
Scalability
Microsoft SQL Server is designed to scale with clients from 5 to
10,000 users. With our diverse range of sized clients we needed a product that
could look after our 1 to 5 users operations through to our plus 50 terminal,
multi-national clients. By using Microsoft SQL Server
or MSDE,
we were able to look after our existing clients as well as be assured that as
new clients join Nimbus and our existing clients grow we have the product that
will cope with this.
Open Architecture
There is major pressure from all involved to have
open systems that can integrate easily with multiple applications. Our own software design
and Microsoft SQL Server has ensured that we have a product that you can easily build your
own applications for, or interface to Nimbus via our Application Programming Interfaces
such as Intranet Web Reporting.
Backups and Transaction Logs
Microsoft SQL Server can be configured to do online backups, for which the time and
frequency can be configured. These back ups can be performed while users are accessing
the system.
For example, a task can be created to complete a full database back up each night.
Additional to this a separate task can be created to backup the transaction log on an
hourly basis. In the event of a system failure the system is restored to the last good
full database backup and the transactions logs are restored sequentially till the last
good transaction log has been restored, thus only losing an incremental amount of the days
processing.
Integrity of Data
Microsoft SQL Server is transaction based which means in the event
there is a power cut or some sort of disruption to your system only complete transactions
will be accepted, thus eliminating data integrity errors when only parts of the
transaction are processed.
|