Is this a viable alternative to CSS11501?

[alec.waters]

Hi,

Assuming the performance was good enough, would the below be a viable alternative to the CSS11501? I'm guessing it's the CSS's hardware architecture that bumps up the price?

http://www.brainforce.com/eng/produkte/ina/ina_b100_cont_eng.htm

http://www.brainforce.com/pdf/b100_install.pdf

thanks,

alec

[Gilles Dufour]

Indeed, the hardware is very important.

CSS does a maximum in hardware for better performance.

[if your loadbalancer is weaker than the servers, there is no need of a loadbalancer].

The company name is also important.

Think about the support for your product.

Take into account the impact of a failure of the product and which company can guarantee less failures and a faster resolution time for a failure.

If you want to compare the CSS11501 with another vendor, please contact a Cisco sale person.

He will explain you the pros and cons of the CSS.

Regards,

Gilles.

[tuan.do]

Hi,

Have you looked at the F5 products. They seems to be quite popular with LB /SSL product.

td

[Gilles Dufour]

http://www.veritest.com/clients/reports/cisco/csm_comparison.pdf

This is an independent report comparing Cisco with different vendors.

Also, as mentioned before, consult one of our sales engineers, he will be able to find a solution for what you need.

Regards,

Gilles.

[ct_yau]

The report you referenced is more pro-Cisco. I found one from the same organization that is more pro-F5:

http://www.veritest.com/clients/reports/f5/f5_performance.pdf

How to understand this? Would you say that F5 is better in the appliances while Cisco is better in the higher-end core switches?

[Gilles Dufour]

You have to look a carefully at these performance tests.

A few remarks

- The F5 documents compare the F5 product only to CSS 11503. The Cisco test was done with the CSM and the CSS 11506 which are higher end platform.

Why F5 did not compare itself to these products ?

- the F5 documents used version 5.20(0.1) for the CSS. This is the worst version we ever had and nobody uses it - would be interesting to see a real version like 7.20 or 7.30

- next the config of the CSS. It uses 'no persistent' which means we have to rebalance every GET request and they did not configure persistence reset remap which will optimize this rebalance.

[was it on purpose ? or they just didn't know it ?]

So, it looks like that with a very old/bad software, with an incorrect config and with very small packet sizes, the CSS is worst than F5.

With the same config and version but large packet size the difference between F5 and the CSS11503 is not that different.

BUT, if we use the 11506 with the right software and the right config, you get the number in the Cisco document and they're much better than F5.

Finally, evenif the performance would be the same, there is all the rest.

Is there such a forum for F5 products ?

Do you have certified engineer answering your questions related to F5 products ?

In case of a software bug [this is a reality, there is bugs with every software F5 and Cisco included], who's going to fix it first ?

I hope this helps.

Regards,

Gilles.