Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

First time network engineer here. I don't have a large budget, so I need to learn all of the skills necessary without pulling in a bunch of vendors. I'm charged with building out a 350port network in a new location. I planned on using a few gigabit switches from cisco small business.

I noticed network in our old locations just has one of the regular 10/100 ports plugged between the switches. They are arranged are daisy chained together, not in a hierarchal fashion.

I will be using vlans to separate a guest wireless network from my corporate network, if that matters.

The new switches have mini-gbic available, but my router has only has gigabit ethernet ports.

What's the best way to set this all up for maximum reliability, redundancy, and speed?The router is an excepted single point of failure, but if one of the switches goes down I don't want my whole network out. Can I use multiple ports between the switches to create redundancy trunks?

Thank you! (from a total noob)


share|improve this question

closed as not a real question by Sven, voretaq7 Jun 20 '13 at 21:31

It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form. For help clarifying this question so that it can be reopened, visit the help center.If this question can be reworded to fit the rules in the help center, please edit the question.

You are not a network engineer. Hire one to consult you on designing this network. – Sven Jun 20 '13 at 21:07
"What's the best way to set this all up for maximum reliability, redundancy, and speed?" - per the Help page here: "Your questions should be reasonably scoped. If you can imagine an entire book that answers your question, you’re asking too much." -- you are asking SF to plan out your topology/etc. for you with very little to go on. Sorry mate, but you should bring in assistance at least initially. – TheCleaner Jun 20 '13 at 21:07
You know what pines my ass? A bunch of 'experts' that are too elite to help someone new get started in the right direction, as if they have sometime to lose by sharing their knowledge. 'Hire an expert' isn't an option, if that wasn't clear enough. Do you not think that I would prefer that if the choice was available? I'm being asked to shat out a rainbow. If OP was "prove how much of a failure you are at life by giving a condescending answer" I would award you both with a 500pt bounty. – Jonathan S. Fisher Jun 20 '13 at 21:36
I'm sorry you're so upset by the responses here. Your question is the SF equivalent of asking, "I have this idea for $KILLER_APP, can you tell me how to develop it? I know some BASIC.". FWIW, The 'Hire an Expert' answer is being discussed over in Meta. Also, please see Why professional capacity and What is a professional. – kce Jun 20 '13 at 23:02
First, who are you and why are you apologizing? Second, if you define 'upset' as utterly confounded by the vanity of the two comments, then yes, it is upsetting. It's not like I asked how to perform heart surgery on someone, we're talking about a computer network. As for the "hey I have a great app idea," if you are a software engineer and you're so far inflated with your own hot air that you can't even listen to a common person try to relate to your profession, then your ego could desperately use a figurative colon cleansing diet of squeezed humility. Try teaching, not admonishing curiosity – Jonathan S. Fisher Jun 21 '13 at 0:26
up vote 5 down vote accepted

You need to consider, in detail, how traffic will flow in your network. Determine the bandwidth requirements for things like access to intranet sites and file servers, as well as internet access. This will guide you toward how you can plan your core infrastructure so that groups of workstations have adequate bandwidth in the directions they will send traffic, and how you can structure your network to accommodate the concentration of traffic toward certain nodes. Generally the structure should be tree-like, with links directly to traffic concentration points wherever possible.

The second thing to consider is the size of broadcast domains. In general, network performance is reduced by excessively large broadcast domains (as DHCP traffic, ARP, some windows networking functions, and a number of other things cause broadcasts, which can have a cumulative effect. You don't want broadcast traffic being flooded out across every link on your network. Limiting broadcast domains is accomplished with subnetting and planning your layer 2 topology; VLANs can help with this, as can simpler strategies such as setting up a small number of switches to serve a set of related workstations on a subnet, and using a router. Keep in mind that routing and DHCP will typically need to be provided for each subnet. Strategies for subnetting range from departmental subnets to a subnet for each floor of a building. 350 devices is definitely enough that you will need to worry about the size of broadcast domains.

Additionally, you should consider which areas of your network have specific security implications. Obviously you will want to have a specific subnet (and probably VLAN) for your guest network, with firewalls set up to prevent it from accessing internal services. Depending on your area of business, you may also want to use a similar strategy to isolate the networks of some business units from others.

Port security is also important. Consider implementing a policy whereby unused network ports are disabled or placed on a VLAN with no connectivity. Also consider restricting each access port to one MAC address for most ports to prevent unauthorized hubs or wireless networks.

You can add redundancy a number of ways. At layer 2, you can leverage spanning-tree to create redundant links between switches (which in your network graph will appear as cycles), so that when one switch or link goes down, there will still be a path for traffic to follow around it. This functionality is distinct from routing. At layer 3, the strategy is similar - ensure that every router has a secondary path to places you expect traffic to be concentrated at (your firewalls and border gateways).

If there is any specific part of this you're wondering about, feel free to start a new question - but if any of that was confusing you should either study a lot (look at the CCNA study materials for instance) or hire someone else to do this.

share|improve this answer
Thank you! Brilliant! Informative! Many things to consider and learn presented here! – Jonathan S. Fisher Jun 20 '13 at 21:38

Not the answer you're looking for? Browse other questions tagged or ask your own question.