Infrastructure
Installing an IT solution, or
in this case a network, often provides several options. The myriad of choice
can usually be contained and understood within a cost versus performance
paradigm. Even if you do not fully understand all of the options, one of
these factors may well decide for you. If money is not an option, then you
simply tell your development team to build the best, and care not for the
cost, however, if you live in the real world with the rest of us, then your
choice will fall much closer to the cost boundary.
Having decided upon the
topology, the architecture, configuration, size and purpose of your network,
then it is time to decide upon the infrastructure, that is, the internal
organisations for data access, data storage and most importantly application
services.
There are basically three ways
in which to organise this infrastructure:
- Thin Client
- Fat Client
- Hybrid
Thin Clients
Thin clients have been deemed
as the future of computing, not just for the business user, but also for the
home user. The thin-client computing environment consists of an application
server, a network, and thin-client devices. The thin client is a simple
terminal or other computing device used to connect to servers where
applications and data are accessed and viewed. Thin client machines are so
called, because they contain no CPU, hard drive or RAM. The system box is
simply a case containing a motherboard, facilitating the connection of
peripherals such as mice, monitors and keyboards. Also, connected to the
motherboard, is a network card, providing the interface for the client to
connect to the to the network. No applications or complex operating systems
are installed on the machines either.
This next generation of
computing enables organisations to reduce costs due to the fact that the
terminal machines are “empty” boxes, costing considerable less than a modern
workstation, thus providing an attractive approach for cutting hardware
costs.
Granted, thin clients pre-empt
the need for a fat server loaded with resources, like several fast CPUs,
masses of storage and plenty of RAM. The server will be responsible for all
processing on the network and support for multiple users, therefore the need
for extra resources. Even though the cost of such a machine is likely to be
high, the money saved on thin clients still enables the purchase of a high
performance server, with money to spare.
Thin clients lower the Total
Cost of Ownership (TCO). As well as costing a fraction of the price to what
we normally pay for a power PC, they are more reliable than traditional
computing environments and require less maintenance. Management, systems
support, and downtime are the fastest growing and least manageable costs to
owning technology. With computing power, applications, and data centralised
on secure servers, fewer technical staff can support many more users. Thin
clients have no need for hard drives, memory, or other hardware upgrades.
Software updates are distributed at once to all clients across the network,
creating a few standard desktop configurations to ease training,
troubleshooting, and maintenance. Server backups secure and protect data.
Fat Clients
As the term suggests, fat
clients are the very opposite of their thin client relatives. Fat Clients,
also known as bloated clients are fully loaded PC's, desktops or laptops,
containing a full suite of PC applications, a PC operating system like
Windows, and network connectivity software. The software resident on these
machines is known as "bloatware". These machines are costly and known to be
complex to manage and introduce the need for more, and better trained
support staff. However, machines of this type actually reduce network
traffic since they run applications and process locally, requiring network
access for print and data services only. Money can be saved on server costs,
but probably not enough to justify a fat client infrastructure.
Hybrid
Which infrastructure is right
for your environment? It would seem that they both are. The fat client will
continue to serve the heavy PC user or the mobile user, and the thin client
will provide PC capabilities to large organisations that are largely
application and data-centric, with perhaps the desire for Internet access.
Again, we are faced with the
cost versus performance paradigm, and it is either of these two factors
which will decide what infrastructure is the correct one. In terms of
objectivity, there is no right and wrong, choices are purely subjective and
are made on requirement or taste. The only requirement, whichever you
choose, must fulfil the purpose for its implementation. |
