June 10th, 2013, 03:26 AM
Switch and NAS redundancy
First of all, thanks for accepting me. This is my first post here. Congratulations on the forums , they look great!.
I am a medium System Administrator, not an expert, and I am just wondering two things about networks:
1- Is there no redundancy for switches?. I have my LAN in a switch, but if it fails...is it my only option to replace the switch?. As far as I know, there is no possible failover cluster here, a hardware redundancy mechanism, as in a Exchange cluster, SQL cluster, etc.
2- We have a NAS with several RAID disks, but still it is only a NAS, so, again, if this fails, what could I do? . Should I consider to buy another NAS for redundancy?
Thanks a lot in advance!
June 10th, 2013, 06:35 PM
It's a cost / benefit analysis:
- How much money do you lose if the hardware goes down
- What is the probability of the hardware going down
- How much money does it cost to have redundant hardware
If the network is critical to the operation of your company then it would make sense to have spare hardware on hand in case of a failure.
I believe that it is possible to have redundant switches in a sense. Not in the same way as a cluster of server works, but by providing redundant pathways between key parts of the network. However, swapping out a single switch really would not take very long at all.
June 11th, 2013, 03:02 AM
Originally Posted by E-Oreo
In my company , nothing can stop working , it is a high availability environment, even when it is not a large-sized company, once we stopped working two day because of a virus and it was terrible.
So, after posting the question I came across "Spanning Tree Protocol" and I saw that that could matches my requirements, it looks great!
As for the NAS redundancy, I still wonder what to do if that NAS breaks down, because in that NAS we have the Exchange and SQL data, as in most NAS. This worries me a lot. We have daily backups of course, but it would take me a long while to recovery all the Exchange and SQL Data, maybe a whole day?.
Also, we only have one router and one ISA Server (Firewall), so, again, if either the router or the machine in which the ISA Server is implemented break down, it is all the same problem, to make a long story short: In Computing there must be always two firewalls, routers, switches, NAS, ....that is the idea of clusters.
June 11th, 2013, 08:20 AM
If you're environment truly can't stop working, then you need to move to a virtualized environment. It's becoming so common place, it's pretty much a "common sense" rule now-a-days with companies that can afford it. You can have a fail-over machine so if the physical box hosting your virtual servers fails, they will automatically move over to the other box.
You can also consider moving certain apps or services to the cloud if they're mission critical. It sounds like you have a small organization, if so, sometimes software-as-a-service (SAAS) can be a good route. Consider moving email to the cloud, etc...
So many options available out there, you just have to do a little research.