Cisco Support Community
Showing results for 
Search instead for 
Did you mean: 

Welcome to Cisco Support Community. We would love to have your feedback.

For an introduction to the new site, click here. And see here for current known issues.

New Member

Management Interface - eth1

Hi all,

I am just curious as to how many users have implemented their MARS appliance with only eth0 attached to their network and how many have implemented using both interfaces where eth0 is for collecting information and eth1 is for management?



New Member

Re: Management Interface - eth1

I have done both and it didnt seem to make a difference.


New Member

Re: Management Interface - eth1

I am curious to know too.


Re: Management Interface - eth1

I have used both interfaces on numerous occassion and it is my preference for two reasons:

1. Putting ethernet 1 on a management VLAN is a Best Practice and makes your box less susceptible to a DOS attack on Ethernet 0.

2. I have seen a noticeable difference in the refresh rate of the Management screen when I use ethernet 1 to access the box. One of the developers of MARS told me they put more memory into the ethernet 1 interface so it would refresh faster.

Just my 2 cents.

Hope this helps.

New Member

Re: Management Interface - eth1

I have a MARS 110r deployed using eth1 for management and eth0 for collecting. This setup works well as others have stated due to the fact that MARS is more responsive while managing the box without adding additional stress to eth0. Previously I had only configured eth0 and used it for both purposes. I shortly found out that I was hindering the performance of the box as well as loosing valuable time to manage the box due to slowness.



Re: Management Interface - eth1

We use a single interface, eth0. It is a gig port and PEAK utilization is < 5% with a MARS 210 processing 175 million events per day. From a performance standpoint, at least conceptually I can think of no reason why connecting on one interface or the other should matter unless the interface itself is the bottle neck. Given the numbers we see, I can't get excited about this as a performance enhancing thing. I'm not sure how this provides much security. The same daemons are listening on both interfaces, the only thing your doing is lighting up another interface and protecting THAT interface, which will only be accessed via HTTPS anyway. Perhaps if you were blocking access to the MARS monitoring interface externally, but then you could just limit management access to specific source networks. I don't want to diminish those that have had success with this, do what works for you. I suppose you could always light up both ports and have them available so that if you start having general slowness you can't test using the other interface.