Need immediate IT support? Even if you're not a customer call 01329 888625 now for priority support!

Instant IT support? Call 01329 888625

How Redundant is your Redundancy?

Posted on 7th April 2017

I’m going to talk to you about ideals here…

In an ideal scenario, where budget is not limited, but, you are still being sensible - what would the best IT system look like for your business?

This scenario won’t apply to every business out there, but you can take what you will from the scenario and still apply this to your own scenario.

I’ll also keep technical talk to a minimum where possible. For those who want more information please do get in touch and we can get as in-depth as you wish!

POWER

A simple UPS (Uninterruptable power supply) Backup Battery will suffice for most companies.

The UPS should be specified to have enough battery to safely shut your IT network down in the event of a power cut, with room to breathe for expansion and covering minor power blips where a full shutdown would just be inconvenient.

30 minutes is usually the standard we work by for our customers.

A single UPS can protect against a power cut, but what if the UPS failed during a power cut?

Then you need 2 UPS units instead.

IT equipment  can be loaded equally across the 2 to ensure a single UPS failure won’t cause an impact to the business.

This can be taken one step further to ensure both UPS units are on separate electrical feeds, to avoid a tripped fuse from taking out the IT systems.

Physical Servers

You should have at least 2 servers in place. Any less and a single server failure can put your business on hold until it can be fixed.

Those 2 server should be ‘virtualised’, which is a clever term that can split a single physical server into any number of virtual servers.
This means that a single server can actually be a Mail Server, Data, Print or Remote server. You don’t need to buy 4 separate servers to
do this anymore.

The cool thing about this is that those virtual servers effectively float across your 2 (or more) physical servers. So if a single physical server has any issues, you can seamlessly move those virtual servers to the remaining, working server.

Servers should themselves be redundant where possible. Dual power supplies, fans, redundant memory and have at least 2 connections to the network in case any one of the above failed.

Virtual Servers

Fantastic, you have a redundant physical server. But what if a software update or bug crashes one of your main Data servers?

You can avoid this issue by introducing redundancy into your virtual servers.

A single data server going offline can grind the company to a halt.

If you build a second data server, which runs from a different physical box (as above) then it reduces the risk of this second server also having the same issue as the first.

Of course there is still a chance of this, but the answer to this? Another data server.. !

You can cluster many servers or services together to reduce the risk.

Storage

Since the servers above are sharing information between them, they need to speak to a centralised device to store all of the data. The operating systems, your email data, your user data.

Typically this would sit on something called a SAN. (Storage Area Network).

The SAN will be redundant in itself with at least 2 of everything; 2 Controllers and 2 Network Connections.

The utmost important thing to specify in your
SAN is the hard disk redundancy. The more disks you buy can either give you more physical storage space, or it can give you more redundancy.

Ideally you should be able to suffer a couple of disk failures without it impacting your network.

If you are monitoring your IT hardware closely then this should be fine, however if you do not have monitoring arrangements in place then you may wish to go as extreme as doubling the amount of storage you have to be able to completely mirror everything a second time.

Switches

Most companies will have redundant servers and storage, but not realise that a single switch means a single point of failure.

By implementing 2 switches, you can equally connect the above hardware and spread the load so if even a single switch fails then the majority of your company can still work.

Internet

With businesses ‘moving to the cloud’ in recent years, Internet connectivity has become a major requirement for your business to operate.

As with all of the recommendations so far, your internet connection can easily be a single point of failure, but can easily be solved.

No, don’t take out a second Internet connection with the same supplier. Instead, take out a second connection either with a separate supplier, or speak to someone like
us who can provide Internet connections
from multiple suppliers, be that BT or Virgin for example.

You can delve into this further and argue that a Fibre broadband connection from any provider will still ultimately connect back to the same BT street box, and you would be correct.

The way to avoid this? Different
technologies - almost like a pick’n’mix!

There is plenty to chose from, at varying price points. Fibre to the cabinet, Fibre to the premises, Dedicated leased lines, Satellite, 4G.

For the larger customers we have even been able to have multiple leased lines installed, and ensure they are run from different locations to ensure resiliency.

Your broadband router also needs to be able to handle these multiple connections, which leads me on to…

Routers

Two internet connections are great. But again, a single router is a single point of failure. If that goes then you have no internet!

You can combine this with multiple Internet connections as above to be able to suffer a router failure, or an internet connection
failure.

Some very clever technology can be used here so two routers appear as a single router on your network, so there is no manual interaction required if one fails.

Buildings

The last one and this is a big one.

If you lost your building in a fire, what would happen with your IT?

There are many ways to protect against this, from buying everything a second time and locating this in a second location - to moving your existing IT equipment to a secure datacenter with less risk of fire, theft or power issues.

Depending on the side of your IT network, this can be a very, very costly option - but an option all the same!

Without going too much into depth in the techie-talk, that’s pretty much it in a nut shell.

Yes there are a few different ways of achieving redundancy, online backup systems and such - but I wanted to concentrate on a standard model that would apply in most situations. This even applies at the ‘cloud services’ level.

If you are looking at hosted cloud solutions then I imagine you will want to be guaranteed 100% uptime, because the service provider should have the redundancy in place.

Chances are this isn’t the reality and it can quickly become a huge issue, since they effectively own you and your data.

Of course, all of this comes at a price and ultimately it will depend on how deep your pockets are as to how much redundancy your business has.

At best case, take a think about how much of a financial impact a single days worth of outage could cost you and then have a think about investing this amount into improving your redundancy. For larger firms it’s worth increasing that figure to 7 days, perhaps a week or two and then doing the same exercise.

Trust me, when it happens you will wish you had the extra redundancy in your IT!