Article

How not to stuff up selecting a server

The server is probably the most important piece of kit on your computer network. Choosing the right one for your school takes planning and know-how, writes Greg Adams.

What is a server?
If you’re not familiar with servers you may think of them as some mystical beast with supernatural powers. While they’re important, the truth is slightly less fanciful. A server is simply a computer whose role is to manage and share resources for other computers on your network. What can make the term server confusing is that it can refer to both hardware (the actual computer) and software (the program running on the computer).

What does a server do in your school?
Servers often have a dedicated function. For example, a ‘file server’ stores files. Any user on your school’s network can store files on it. You may have your financial systems, LMS, firewall, and back-up files on a server. A print server manages printers that all users can access (as opposed to having a printer connected to each computer in your school). A network server manages network traffic, and so on.

Choosing a server
If you’re the one choosing your server, make sure you’ve done your homework. It may not be as simple as adding server software to a PC. Your choice is as much about understanding the consequences of a wrong decision, and making the right one. If you’re not confident in your abilities, don’t be afraid to ask for assistance from experts. They know the gear, they know what servers can do…  and, importantly, what they can’t. Provide them with the information they need – and enough for them to make the right choice – and they’ll come up with options.

There are a number of factors involved when deciding upon the right server for you, including such things as:

– Environment – in what physical conditions will it operate?
– Redundancy – what happens if it fails?
– Data storage capacity – now much memory do you need?
– Data retrieval
– Number of users – who’s logging in and when?
– Security
– Cost – what are you getting for your money?

It’s easy to focus on the piece of kit and forget about the environment in which it will sit. However, this is arguably the key aspect you need to think about. Will it be housed in a dedicated room… or sat on the floor by the office printer? A purpose built rack provides protection and better security (most are lockable).

How reliable is the power supply to your school? If you’re a rural school that often experience outages – or even if the lights just flicker from time to time – your approach will be different to that of a school in one of the main centres. You may need an uninterruptible power supply (UPS), also known as a battery back-up, to provide emergency power. Try to make the environment around your server as good as you can and you will minimise the potential for problem and failures.

You’ll need to think about how important it is for you to have your system running. Can you cope with a few hours out or do you need it working constantly? Do you need a second server for redundancy? Whatever you decide, any equipment configuration and service level agreement will need to achieve that.

Another important aspect to selecting a file server is storage capacity. Ensuring you have enough space for current requirements and future growth is essential. It’s also very important to make sure you have the right backup solution that is capable of backing up the entire server, on a regular basis.

Servers tend to last longer than general PCs – as a rule of thumb up to five years instead of three – therefore, consider the lifespan, both in terms of your planning and warranty protection. Cost, of course, will always be on the radar. Servers start at about $1,000 and go up from there depending on your requirements – from ‘entry level’, through ‘performance’, to ‘optimal’ servers. If your determining factor is ‘best price’ look carefully at what you’re not getting for your money. For example, 2GB DIMs are cheaper but may quickly use up your slots; the more expensive 4GB DIMs offer more storage and, generally, consume less power.

What is virtualisation?
Virtualisation is a long word for what seems a simple enough concept. It’s a piece of software that allows core computer systems to be run on a single server – systems that traditionally tend to run on separate servers. Products like VMware and Hyper-V. Don’t worry too much about the ‘how and why’ of the technology. It’s what this virtual ‘pooling’ of resources can achieve that’s the interesting bit.
Supporters of virtualisation put forward a number of potential benefits of adopting this way of doing things, these include:

Less hardware – programs sharing the same space means less hardware is required to run the same applications.
Improved performance – with a virtual system able to allocate server resources where they’re needed, it has the ability to respond if applications require more memory, processing power, etc.
Lower maintenance costs – consolidation can reduce system management as there are fewer pieces of hardware to maintain.
Disaster recovery – as resources can be ‘moved’ around in virtual environments, there is a greater ability for recovery should a server fail.
Better for the environment – fewer servers means less power consumption, less to throw away come upgrade time, and less air conditioning.

However, that said, virtualisation is not without its challenges. Ironically, what makes it attractive can be its downfall – namely that it sounds so easy. Of course, it’s not. Unless you really know what you’re doing, virtual servers are probably best left to the experts. By their very nature, virtual servers run harder than discrete servers housing only a few applications. These systems require fairly high-spec hardware, which does cost more – perhaps twice as much. There’s potentially ‘more eggs in one basket’ when it comes to redundancy. Also, virtualisation doesn’t work for everything. Firewalls, for example, need to be on a separate server for security reasons.

Avoiding common mistakes
We asked server providers to list some of the common mistakes they see schools make.

They replied:
– Thinking about a box and not an architecture;
– Buying an under spec’d box – using entry level server when needs (present or future needs) are much greater;
– Not thinking about future needs;
– Selecting the wrong server software for the environment;
– Schools still store important data on workstations rather than on the server;
– Not having a suitable back-up solution in place; and
– Forgetting to swap the back-up media or believing the server has some magical automated process that swaps the backup media for them.

Future developments
As for what’s going to happen to server technology in the next 12 to 24 months, there are a number of things you should keep an eye open for.

An on-going development is what’s known as cloud computing. Here server functions are provided as an online service – in other words, the server is in a different physical location and you access what you need over the Internet. As networks improve, this may become an option for your school, albeit with the usual caveats about security, confidentiality of data, reliability of links, and so on. Microsoft’s new operating system Windows 7 is promising better network management than XP with Server 2008.

Expect the hardware to continue to evolve. As technology matures, six core processors (from quad/four) will start to appear in lower class servers, and new servers offer space for 12 disks, as opposed to the six of previous generations – so consider buying boxes that can be upgraded rather than buying the whole lot up front. New chipsets will appear that are faster, cheaper and more power efficient. Solid state drive storage is coming down in price and will become more prevalent, and 10GB networking is also becoming more affordable.

Whatever server you choose it needs to fit your school’s ICT strategy and not be a reactive replacement decision. Thought and planning needs to be given to continuity (what happens if the server needs to be off line for maintenance) and disaster recovery (how is the system recovered if the hardware fails). You need to think as much about the system architecture as the box.
If the product range seems huge, that’s because it is. You need to be honest about what you require so that you – or your provider – can make the right choice. The correct decision will pay dividends during the life of the system through lower management time and lower system down-time.

You need a plan and to know what you’re doing. Selecting a server is a decision that should not be taken lightly.

Greg Adams is editor of INTERFACE Magazine.

Copyright G Media Publishing Ltd. 2014. All rights reserved. Privacy

Categories: Article, Issue 20