FUTURE COSTS (part 1)

From the developed model to the actual model, costs continue. There are several factors that continue to make demands of financial resources.   Upgrades  The IT industry is dominated by change. No sooner are new, cutting-edge technologies installed, it seems they are out of date. Technology and growth move at a staggering rate. In the High Street, as the higher-purchase consumer makes the final payment on equipment purchased a scant 30 months ago, knows fine and well that the equipment is already antiquated. This of course affects not just High Street consumers, but businesses also.   Money lavished on development costs and system implementation, is not the end of the matter, indeed, this is just the beginning. Granted, it is highly unlikely that such large sums of money will be spent on one-off purchases again, but ongoing costs are intrinsic to the nature of this work.   The table below is a summary of some situations that incur the need for upgrade at one point or another. Situation Solution More traffic to Web server  Install faster access media (ADSL etc) Upgrade RAM, CPU, Hard disks Increased security risk Install Firewall/Proxy Server Upgrade to more secure Firewall software Upgrade to more complex monitoring software More Local users Upgrade network by: Installing hubs switches, bridges and routers  Adding new workstations Installing additional server(s) Install more peripherals Performance Slump Faster network cards Segment and Subnet (Switches and bridges)  Upgrade to faster cable Upgrade to faster components (Switches and hubs) New Software Incompatibilities Upgrade Operating systems Upgrade Applications Software Upgrade hardware New Hardware Incompatibilities Upgrade Software for smoother integration Upgrade Existing Hardware    Upgrading can be kept at bay for a time, limited for another time, but can never be avoided. When installations are put in place, it is advisable, if at all possible to provide more than is required. This way, your application will deal with growth without incurring immediate costs. As use increases, the system can cope because of the measured over-compensation, wisely inbuilt at the design stage.   Inevitably, the need will arise for new components, and this is where a well-designed system can stave off higher costs for a time. Building your system with the ability to expand, will lessen the amount of immediate spending, preventing it from being too costly in the early stages. Applying some savvy at the design stage, will allow upgrading to be carried out without having to replace core components. For example, installing hubs on the network with spare ports, will allow network expansion without cost (apart from the new node!). If there are no spare ports, then a new hub is required. Use latest software releases to prevent hardware and software incompatibilities at an early stage. Use the fastest technologies affordable to avoid wholesale upgrades. The basic rule here is more of everything as budge. Development Costs  Building and hosting a Web site is not going to be cheap. The intrinsic nature of IT is one of great expensive. Granted, it is possible to purchase cheaper equipment, however, it is either a brand of ill repute, out of date, or both. Equipment of this type, must be avoided, otherwise problems are likely to reveal themselves further down the line in time. Hardware  The server machine will be the most expensive item that you will have to purchase. Its cost will vary, depending on its configuration. The sites functionality will determine to a large degree the choice of components to be made, and the way in which they will be configured. If you are running a mission-critical application, then you will need to buy the fastest, and the most resilient available.  You may wish to consider the following for your system: Multiple CPUs  Since the CPU is able to process only one instruction at a time, then it will make sense to install perhaps two or even four, depending on the server load. The CPU is the most expensive part of any computer system, so multiple CPUs will drive costs up. RAID (redundant array of inexpensive disks)   RAID is a technique that uses multiple hard disk drives. When a file is saved, it is striped across the disks, with parity information being created also. Should one of the disks fail, the technician is able to quickly swap the disk out without having to down the server. Once the new disk is in place, the parity information on the other disks will rebuild the lost information on the damaged disks. Even though the cost of storage is decreasing, multiple disks still incur greater costs. FUTURE COSTS (SCSI) SCSI (small computer systems interface)  This is an interface standard, required to connect devices to the system. SCSI hard disk drives are used in servers because they are faster and perform better. Of course faster means costlier.  RAM (Random Access Memory)  The greater the amount of RAM the better. RAM is not anywhere near as expensive as it used to be, and a gigabyte of memory is very affordable these days.  Motherboard  The main system circuit-board, which facilities the connection of all devices to the system. Using the components mentioned above, by nature drives up the cost of the motherboard. A motherboard with at least four SCSI controllers, and two CPU sockets, will be required, these do not come cheap. Some motherboard manufacturers build boards with a propriety bus and multiple CPU sockets, which further increases costs.  Backup  Statistically speaking, the hard disk is the component most likely to fail. RAID cannot account for fire or flooding, and is therefore limited in what it can do. This is why backup is essential. Should disaster strike, having backup copies of the hard disk drive will allow you to reinstate the server. There are many different backup drives and techniques available, these can be internal or external to the server. External drives tend to cost more, due to the extra casing. UPS (Uninterruptible Power Supply)  This is an essential component to keep your server running if the power fails. Power spikes, brown outs and black outs can be very damaging to your sensitive electrical equipment. Fluctuations in the power supply can cause memory and hard disk errors, and black outs can totally corrupt the operating system. The UPS is like a battery pack. It is plugged into the main power supply, and the server is then plugged into the UPS. The UPS then takes care of drops and surges in power, providing a smooth electrical current to the server's sensitive components. In situations of total power failure, the UPS will hold charge for a certain amount of time, until either the power is resumed, or the administrator is able to get to the server and close it down gracefully, to prevent data corruption. UPS vary in price, the more expensive they are, the better they handle rogue power outlets and will hold charge for longer in power-failure situations.  Below, is an excerpt from the Dan Web site, that the price of a typical server. Many of the options mentioned above are not listed, so the price will likely rise by at least another £1000 perhaps double, or even treble that. Software  Software is not cheap either. Unfortunately, there are many software components required, again driving up costs. You will most likely need most of the following. Server Operating System The core of your Web site is obviously the server machine, therefore it required a stable and reputable operating system that will not fall over under increased stress. As mentioned in other modules, the choice comes down to two types, Unix or NT. These days, more and more people are opting for NT (now Server 2000), simply because it is easier to operate, maintain and configure. The server machine comes with Server 2000 per installed, but is likely to cost around £1000 if you but it separately. Web Server Software This software is required to run in coalition with the server operating system, and provides Web services on your machine across the Internet. There are many free solutions available from the Internet, that are often better than the commercial versions. NT users need not worry about cost in this department, since they get fully functional Web server software integrated with the NT software. It pays to research and test the various freebies, before throwing money at software, which may be more than you need. Below is an example of a commercial solution, $995. per licence!!! Development Applications  It is possible to build quit elaborate and functional Websites from basic tools. In fact, many designers will still use the Windows Notepad and browser software to design Web pages, however, most professionals enjoy the functionality and ease of design tool software. These can be purchased individually or in suites. Macromedia is one such company who provide a suite of applications to aid With Web design. These applications automate the design process, with the users not really having to know a great deal about things like HTML JavaScript etc.  An example is shown below. www.micromedia.com Testing Software  As mentioned in previous units, software is required to perform load balancing and monitoring. Again, Windows NT users enjoy the pleasure of having integrated monitoring tools for "free", however, load-balancing software is very expensive. It may actually be advantageous to enlist the services of an external contractor to test your system for you. Alternatively, be prepared to pay big money if you decide to buy your own. Staff Training  One expense often over looked in IT, is staff training. Training is a much-needed provision. There is little point in deploying state-of-the-arts systems when no one knows how to use it or maintain it. Keeping staff on the ball, will compliment your IT strategy. Training can be provided in-house, which will require the services of a dedicated team with a broad skills and knowledge base. Often the services of a third party training team can save money, since they can be used as and when required.  Major players in the IT market have begun to set up creditable qualifications that are becoming widely known and accepted in industry.   Microsoft has a whole range of different modules available to individuals and companies. These modules can be acquired and built into group awards, like Microsoft Certified Professional, Microsoft Certified System Engineers, Microsoft Certified Systems Administrator etc. Training of this type normally assumes the form of 3-day and 5-day courses, followed by a multiple choice test taken at a later date in a Microsoft approved test centre. These courses normally cast around a minimum of £1000 pounds each, however, it is possible to buy a Microsoft approved textbook, study the material, and apply for the test when you are ready to take it. The tests cost around £100 each. Cisco is another major player who has joined the training market. They provide training mainly networking, and so their qualification is extremely advantageous to individuals working in the Web server branch of IT. The CCNA (Cisco Certified Network Associate) involves a web-based teaching approach, with an in-class tutor to help with enquires, set-up on-line tests and oversee practical tasks. Acceptance Trials & Pilot Schemes Part of the development budget, will involve acceptance trials and pilot schemes. Acceptance trials will likely involve installing a prototype solution that will allow the organisation to test their ideas and modify them as required. In other words, should an idea fail, it is either discarded, or improved. The newly improved idea is then subjected to the same testing regime until it functions sufficiently well. The new server machine can be purchased and run alongside the existing system, until it is safe to migrate to the new system. .  Alternatively, an organisation may adopt a pilot scheme of sorts. This will involve building a system entirely separate from the existing system, to serve as a tentative model for future experiment or development. Like before, this system would be deployed and monitored, often modified and re-assessed, until acceptable results are acquired. The model may well remain after a satisfactory conclusion is reached, and an entirely new system purchased and deployed. The testing system may well remain as a tool for future development, either for entirely new systems or to test modifications to the existing system. Data Conversion It should be noted, that the deployment of any new system, is very likely to impact applications and/or DATA, causing them to function undesirably on the new model. This depends upon the extent of the change. If the new system is the same in principle to the older one, i.e. moving from Windows NT to Windows Advanced Server 2000, then perhaps all that will be required are upgrades in the applications. However, if the new system is an entirely different one, with cross-platform data formats, i.e. Unix to Windows, then all previous applications will not run on the new system, and a new suite of applications will must be purchased. On the other hand, there are two options with data, either hard copy formats are maintained and then painstakingly re-entered, or it may actually be possible to convert the cross-platform data to the new formats. There are companies who offer these services for a fee.