rb|it : blog

SME Disaster Recovery - Server Resilience

Your server is made up of multiple components, any one of which could possibly fail.  If you’ve ever found your company suffering because “the server’s down again” then something is probably wrong with the way it was setup.  Any server, even a low spec one, should have as few ‘Single Points of Failure’ as possible (i.e., as few places as possible where only one ‘widget’ is doing the job, and failure of that ‘widget’ brings the whole server down).

Here’s how to review your server components and make sure you’re protected against single component failure.  You don’t need to be a technical guru to check these things, the quickest way is to examine your original purchase records to check the specification of your server, if in doubt, ask the original supplier or the person who set it up.

1. Disks.  Disks are one of the most common points of failure on a server and one disk failing should NOT bring your server down.  You have choices on how to have the disks setup, but they all revolve around something called RAID (see http://en.wikipedia.org/wiki/RAID).  In a nutshell, you duplicate disks so they share the work, and if one goes down, everything keeps running.  There are two choices of disk type, SATA (cheaper) and SAS (faster), and there are two main types of RAID configuration; RAID1 (initially cheaper but becomes expensive as you increase your storage capacity), and RAID5 (initially more expensive, but cheaper for more disk space as you don’t need to buy as many disks).  All your servers drives should be provided using RAID so if one disk fails, your server keeps running.  Even better is ‘hot pluggable’ disks, which are much easier to replace (with care almost anyone can do it in five minutes), they just slot in, and slot out, and you can do this while the server is running.
2. Memory.  If you have only one memory module, then that module is a ‘Single Point of Failure’.  If it dies, so does the server.  A better option is to use two, or even four memory modules.  Ideally if one module fails you should still have the minimum required to run all server functions, so pad a bit (i.e. if your server runs OK with 2GB, buy 2 x 2GB modules, it will probably run faster, and if one fails it will still keep going).
3. Network card.  Most low end servers come with a network card built in, it’s wise to supplement this with another network card.  You don’t need to set up any technically complex or expensive configuration, it just means that if one fails, you (or your technical support provider) should be able to quickly configure it to replace the failed component, instead of waiting a day for a new network card to be delivered by your supplier.
4. Power Supply Unit (PSU).  The PSU failing will instantly bring your server down.  Add a second PSU which runs simultaneously with the first, so a PSU failure doesn’t even slow the server down.

Costs

You might think this ‘big business IT’ stuff costs a lot of money, but it doesn’t.  Below are the costs¹ of implementing the redundancy measures outlined above on an HP Proliant ML350 G5 server (which purchased with one disk would cost £858 ex VAT).

1. Disks.  Adding 3 more disks to allow you to implement RAID1 would cost £221
2. Memory.  The server comes with one 1GB memory module.  Upgrading this to 4GB (2 x 2GB modules) would cost £120
3. Network card.  The server comes with one network adapter, adding a second adapter would cost £29
4. PSU.  The server comes with one redundant power supply, adding a redundant PSU to an ML350 G5 would cost £104

The total cost of all these options is £474, making the total server cost £ 1,332.

Think that’s too expensive?  Offset it against the cost of downtime.    Taking a model company² of 20 staff, the cost in salary alone (i.e. the percentage of computer-related productivity lost in terms of employee salary) is over £900 per day!  This doesn’t include any lost sales, the negative impact on customer relationships, the cost of the time taken to recover from the outage once it is over etc.

If you’ve had one outage in the last year, it’s well worth looking at these practices, because they will save you money, keep you more competitive, and possible make the difference between a good year, and a bad year.

Richard Bartlett offers expert and independent IT Consultancy Services to review your IT, help buy the right systems, install them, train your users to use them, and ensure you have appropriate IT Support. For more information on services including installation and configuration services which can give your systems resilience see http://www.rgbartlett.co.uk/services/

¹  All prices taken from ebuyer.com on 28th August 2008.
²  Based on a model using average salaries for the Cambridgeshire region supplier by Reed, and an algorithm developed to calculate downtime costs.  If you would like to know the cost of downtime for your business get in contact (see http://www.rgbartlett.co.uk/contact/).

Disaster Recovery for Small Businesses

In a typical small business, say 15-20 employees using one server, if that server goes down it's a disaster.  Time lost, sales lost, the cost of IT Consultancy to rebuild and recover data, it all adds up.  At a time when remaining competitive is critical, small businesses can ill afford to be off-line

Traditionally, Disaster Recovery (or 'DR') has always been Large Enterprise territory, but not all is lost.
Whilst full site DR is beyond the budget of small businesses, you can build in a huge amount of what's called 'redundancy' into your systems, to the point where total component or service failure would become an inconvenience, instead of a total disaster.

Over the next two weeks I'll be looking into various redundancy options which are available;

• Reducing single points of failure in your server with components
• Eliminating your internet connection as a single point of failure
• Recovering from a mail server outage quickly on a low budget
• Keeping your network backbone up with redundant switches

All of these options can be implemented at low or very low cost, and you'd probably find you'd have a return on your investment in the first year.  It could even save your business.

Check back at the end of the week for the first post on reducing single points of failure in your server.

Email getting on top of you?

Email is a hugely powerful tool, it allows communication across the world in seconds between individuals or groups, in the form of text, pictures, video or audio material, and it is becoming almost omnipresent, like television.

However, such is our reliance on email in business that it can sometimes become a burden, not an asset.  It's common for people to have hundreds of unread emails in their inbox, in Customer Service related departments in particular, staff cannot cope with the flow of requests and information they receive, and so customers lose out and staff stress levels rocket.

There are a number of Customer Relationship Management (CRM) tools out there to manage this specific problem of customer > supplier communication, but there are still countless other functions in business where emails are overwhelming the recipient to the point where they become useless.

There is another way though.  Two years ago, having reached the point where I could no longer cope with the flood of email into my inbox, I was told about a work management system called 'Getting Things Done'.  The trite title belies the effectiveness and almost unbelievable simplicity of the system, and revolves around keeping your inbox empty (I know, a pipe dream isn't it!).

In summary, it works by;

• Capturing all your inputs (email, paper in tray, phone calls),
• Defining actionable things discretely into outcomes and concrete steps,
• Organising reminders and information in categories based on how and when you need them, and
• Reviewing everything in a systematic way.

There's a book on the method which is probably quite hard work but if you're that overloaded you'll be glad you read it!  There's also a PDF on how to make the most of Outlook (2003 or 2007) available from the GTD website, and it turns this simple application you curse at every day into an incredibly effective personal workflow and productivity tool.

Don't pay hundreds or thousands of pounds on 'specialist' producitivity applications, it's all there in Outlook just waiting to be used.  Set some time aside, and learn this new method, and you'll be glad you did, I guarantee it (speaking as someone with an empty inbox).

Richard Bartlett offers expert and independent IT Consultancy Services to review your IT, help buy the right systems, install them, train your users to use them, and ensure you have appropriate IT Support. For more information on services including training on using Outlook for personal workflow and productivity see http://www.rgbartlett.co.uk/services

Sources: GTD and Outlook 2007 whitepaper © David Allen & Co.

Training: the missing link

Training comes off second best in IT Budgeting and Planning across companies of all sizes, for a number of reasons;  in large companies the budget responsibility falls into the gap between HR and IT, and is first to be cut when times are hard, and in smaller companies staff time away from the office and cost are major inhibitors.  In small companies in particular though, a lack of IT Training at all levels is becoming a significant factor in performance and competitiveness.

At the top, business managers need to improve their awareness of IT, so they can make informed decisions about IT, improve their selection of services and solutions, and recognise what their business needs.  At the coalface, people need training that isn’t just vendor based “how to” training, but training which fills the gap between the basic use of the application and the most effective use for their business.

The gap is there to be filled, and even in this financially difficult climate, companies should seriously consider investing in training, as it could give them the edge they need to survive the downturn.  A recent survey identified "better use of IT" as 80% of the productivity advantage of US multinationals, and concluded that "UK companies assimilate and exploit IT more slowly than their US counterparts, delaying returns".  It's not just training companies that recognise the consequences of a lack of training, in the same survey;

• 92% of the employers not providing training believed the lack of training had an effect on their business
• 43% believed lack of training had a negative effect on product or service development within their organisation,
• 31% believed that lack of training was detrimental to productivity.

A typical case is found in a study by an email managament consultancy, which found that the overwhelming majority said neither they nor their managers have been offered any email training. This lack of training in a critical system disrupts communication rather than enable it,  leading to problems like these;

• "more than 80 percent of respondents think others do not use the email ‘cc’ function properly"
• "62 percent feel most emails are too long and should be kept to a maximum of four or five lines"

Add these inefficiencies together, and what you're doing is slowing your organisation down, and ultimately wasting money.  It's time to fill in the missing link, and train your staff.

Richard Bartlett offers expert and independent IT Consultancy Services to review your IT, help buy the right systems, install them, train your users to use them, and ensure you have appropriate IT Support. For more information on services including training see http://www.rgbartlett.co.uk/services

Sources: Technology Counts: IT and Telecoms Insights 2008 (e-skills UK), "Lack of email training is costing employers dearly" (sme web, mesmo consulting)