You might not be old enough to remember the highly effective marketing campaign from IBM, that implied buying from anyone else would get you fired...
However, I just read Kemp Technologies mildly amusing press release about replacing F5, and it got me thinking -
Is it time that F5 starts to run a similar campaign?

Let's be very clear, replacing an F5 deployment with either Kemp Technologies or Loadbalancer.org could be the worst decision you ever make.

It is highly likely to be a terrible decision if you spend 45 minutes doing the replacement as the latest Kemp press release implies.
And whoever does something impulsive, daft and reckless like that - definitely should be fired.

5 reasons NOT to change your application delivery solution from F5 to either Kemp or Loadbalancer.org:

  • Because it's cheaper.
  • Because the sales guy said it would be easy.
  • Because the marketing material said it was a direct replacement.
  • Because the CFO heard load balancing is now a commodity product.
  • Because the cheaper solution already proved itself on a couple of recent projects.

As I have argued consistently for a long time, F5 is the market leader for a very good reason - It's simply the best thing you can get.
Like most large enterprises - you are running all of your applications through a pair of F5s on the edge of your network and if it works for you, then:

Why would you screw with a great solution just to save a few bucks?
Because you are solving the wrong problem - that's why.

A pair of F5s is a great solution, but mainly if you need to run a crazy number of applications and servers through a single device. In that case, you definitely need an expensive, complicated, over-specified and over-engineered single point of failure. The F5 business model is to get its customers - yes you - totally dependent on a single vendor for all of your edge of network communications.
And considering they earned over $2 billion last year, we think they are pretty awesome at it.

As a side note - Aaron wrote a great blog a while back about how some F5 migrations can end up like a giant game of Jenga.

So what should you be considering doing?
What you should be thinking about is: how to design and implement a simple, cost effective and scalable application delivery solution which is - exactly the right specification for your business - not F5s.

5 reasons you SHOULD think about changing or replacing your F5 deployment:

  • Because it's just getting too complicated.
  • Because you are too scared to touch it.
  • Because you're not quite sure why it was designed like that in the first place.
  • Because you are running out of CPU power for the SSL requirements.
  • Because all my recent applications are in the cloud now anyway.

Obviously, If you are affected by any of the above 5 reasons, then replacing the F5s with another vendor is not going to help you.

If you could start again - What would you make it look like?
No one person or company can tell you the perfect design, and no vendor should be holding you to ransom either.
I do have some suggestions though, I've put them as questions for you to think about:

What if each application had its own secure pair of load balancers?
They would have to be software and/or cheap...
What if the application was designed to scale in the first place?
Think Exchange 2016 rather than Exchange 2007...

What if we ditched all of the crazy DMZs and VLANS and just had a flat network?
Everything would still need to be secure...

What if low level IT staff could use the load balancer safely?
It would have to be very easy to use...

What if SSL offloading, security and authentication was handled securely at the application layer?
Then I wouldn't need to worry about growth...

What if I didn't need application delivery at the edge of the network at all?
Then it wouldn't be my problem anymore...

When you're ready for change, we'll be happy to help - but be prepared though, we will ask lots of questions to ensure you're making the right decision and not an impulsive one based on a press release that implied it only takes 45 minutes to replace an F5.

Feel free to take a look at our full-featured free trial - which also includes unlimited access to our 24/7 support.