FluidOne Blog

Houston, we have a (WAN) problem - The Top 13 problems with WANs

Written by SAS | 25/11/16 15:17

A long time ago, WANs were simple. The problems were simple and managing them was simple. Like the Duke of York's men, they were either up or they were down.  

 These days there are more problems to deal with.  They're more complex, and often they're not easy to measure or manage.

We did a poll of the most most painful and annoying WAN problems for IT Teams. Here are the top 13 that we hear from our customers:

Application performance

  • My application performance is poor and I don't know why
  • My web-delivered application (or website) is slow
  • I don’t know what’s going on with my network
  • Citrix is not working (or Voice, or any other time-sensitive application)

Circuits and Circuit delivery

  • I need a circuit urgently
  • I regularly need temporary circuits for new or temporary sites
  • Generally, it's taking way too long to install WAN circuits
  • I don't know whether my network is fit for purpose
    (MPLS vs Ethernet vs Hybrid network)

Service experience

  • I never get an accurate bill from my carrier
  • It's hard work to manage my carrier (or my carrier is changing things)
  • It’s extremely difficult to make changes to my network
  • I don’t know what my inventory is (I don’t know what my assets are)
  • I want to reduce the cost of my network

Is there a take-out from these problems? 

What's interesting is how few of these problems relate to the traditional concerns about a network. Also, if we're looking to carriers to fix these problems, we're often looking in the wrong place.

Here are three take-outs for you to consider:

"It's the network, stupid!"  Actually, no, not any more.

When we had application problems we used to blame the network and then throw more bandwidth, perhaps more computing power, at the problem.  But things have changed, because increasingly digitalised businesses have more applications, and because cloud-based, web-delivered applications are more complex and more reliant on the network. Our research shows that 31% of mid-market businesses believe they have application performance issues. 

But, interestingly, when we analysed hundreds of performance-related tickets at our Network Operations Centre, we found only 30% were caused by the network.  The rest were caused by server and storage infrastructure (30%) and by applications (40%). We found a villanous but treatable range of causes and we now use the learning to help diagnose and fix performance problems.  

A key point is that to get great end-user experience you need great performance across network, infrastructure and applications.  To achieve this, you need to monitor and manage them together.  

That is where many people find problems.

Their network monitoring either doesn't capture all the components required, doesn't bring them together to show the end to end application path, or isn't backed up by a multi-disciplinary support team.

Bill Clinton's famous war cry needs updating for WAN performance: "It's the application path, stupid"!  

Next time you have an intractable performance problem, get someone to show you how Critical Path Monitoring and application troubleshooting techniques can recognise issues quickly and uncover complex, often unexpected root causes. An IT Managed service might help you with these.

Don't expect circuit delivery times to improve. If you're in a hurry, find an alternative technology.

A typical example would be if you're using an IP Connect network, then a new EAD* circuit will set you back (literally) 87 working days - that being the published lead time the last time I looked.  It's hovered around that for as long as I can recall.

We've seen a huge increase in people turning to bonded 4G LTE cellular WAN to get sites installed quickly. Cellular WAN technology can provide connectivity within 48 hours of order.

Bonding multiple SIMs from multiple carriers delivers reliable bandwidth for multiple users and real-time applications. We recently analysed a batch of over a hundred such sites. We found nearly 90% had more then 3 SIMs and nearly 30% had more than 6 SIMs (look out for a post on this soon).  

There are a lot of use cases for the rapid deployment of new wan sites (or even just rapid deployment of internet).  For example, there's expediting a circuit that's dragged on more than expected (very common). There's repeatable, short-notice, deployment where the IT team commonly gets little notice of new sites. And there's short-term deployment where they're not in place for long.

There are scores of options and design criteria to get right with these solutions, but when you do, you can wave goodbye to many circuit delivery problems. Have a look at what goes in the box, and one example of where it's used.

Finally, to prove the adage that Prior Planning Prevents P!*S Poor Performance, you can now get planners on site before placing an order, to anticipate and reduce some of that delay (there's a post coming soon on this, too!)

Carriers...can't live with them...can't shoot them, either.

Performance issues get the headlines, but it's the soft issues that ruin your life.  

These days you need the network to do your bidding, not the other way around. If you've had your network for some time, you may have started to accept some of the pain that comes from living with it. But time has moved on and new challenges bring an increased requirement to manage the carrier.

Some carriers are great. Some want to be great, but just can't help themselves.  

Perhaps you never get a bill that's right.  Perhaps you struggle with making changes, expediting orders, escalating issues, requesting ceases, keeping up with the quarterly changes to your account manager. Perhaps you'd hoped your carrier would monitor the bits of your network it didn't provide for you. Perhaps your price is not going down in line with the falling cost of fibre ethernet circuits - in which case you should really read our guide to reducing the cost of your WAN.

GET THE GUIDE

 

The good news is that all these issues can be managedThe bad news is that all these issues have to be managed by someone.

It doesn't need to be hard, just organised. This means measured, documented and in one place.

Here are a few thoughts to get you started: 

  • If you haven't already got someone doing this, find someone diligent who can manage the contract, or at least the day to day measurement of what's going on.

  • Dig out the critical paperwork for your network
    • Contracts
    • SLAs
    • Bills
    • Any other documentation, such as your Customer Handbook

  • Baseline your WAN inventory
     
    • You'll need a list of sites, circuits, hardware and management services if you're going to manage them.

  • Find out how to monitor your network performance and availability

    • Perhaps there is a report
    • Perhaps a portal  
    • If not, ask your carrier to provide them.

  • Start tracking and recording your performance
     
    • Use your reports and portal.  Try to monitor trends.
    • Also gather input from your network and IT managers.

  • If you don't already have regular reviews, ask your Carrier to hold them with you

    • Use your newly aggregated data to question the service you're getting.
    • In particular, get proactive about your bills.  Here are some ideas!
    • Hold people to account!
 

In summary

Get someone on the job and stand up for your rights.  

If you think you'll struggle to do this, then consider getting a management service to do it for you. It could save you a lot of time and make you feel far less lonely the next time something goes wrong at 2am. It might also pay for itself just by sorting out your billing.

Finally...

For some, the WAN is the beating heart of their IT. For others, the WAN is a stick that beats the heart out of them.  Either way, it's only going to get more important as you add more applications to your business, so it pays to get to grips with it.  

Also, it pays to get someone to measure your WAN properly, because, like the Grand old Duke of York's men, these days the problem is often that it's neither up nor down.

 

*Acronym alert!


EAD stands for Ethernet Access Direct.  It's an Openreach product that provides the connection between your premises and your carrier's network when you've asked the Carrier for 100Meg or Gigabit speeds at one of your sites.  It's based on a fibre optic cable, presented to you as Ethernet. Alternatives include EFM (Ethernet in the First Mile), DSL (broadband technology, to you and me) and Superfast.