Well, lets cover the options for a small WAN first, then we can get into pro's and con's of each. You could do IPSec tunnels at each site and either use static routes or a dynamic protocol between then to get the information back and forth. I am not sure of what network resources reside in your metropolitan area, but if the sites are close together, you might be able to hop onto a MAN (Metropolitan Area Network). You could use Frame Relay in a hub and spoke deployment. You could use MPLS based WAN services.
Let's get a few basics covered first ... routing. You will need a router at each site, and I would recommend a dynamic routing protocol: RIP, OSPF, EIGRP - something like that. This is so each site can get to each other site and if you ever need to add new sites, or new segments at old sites, the information propogates quickly and automatically. Hub and spoke refers to a design were you have one central site that each remote site connects into. If a remote site (spoke) needs to talk to another remote site, they do so through the hub site. Obviously that is a con if you have much spoke-to-spoke traffic, but it sounds unlikely in your case. The pro is it is the cheaper way to deploy a WAN (the other option is full mesh).
IPSec - This means buying an internet connection at each site, and a firewall or a router with firewall features. You then set up a secure tunnel across the Internet to the hub site. The pro is that this is as cheap as it gets really, and it works fine. The cons are two fold: complexity and scalability. You need to learn how to build IPSec tunnels, maintain and troubleshoot them. If you add dynamic routing on top of them (IPSec in GRE) then the complexity goes up another notch. As you add more sites, you will find that this does not scale well - but for 6 sites? Should work fine.
MAN - These are great options here in Los Angeles. Usually you are buying fiber from a local provider, and plugging it into your switch. It could be a Layer 2 or Layer 3 connection (in my experience Layer 3 is more common). They tend to cost a little more than an Internet circuit, but not too expensive, and much faster and more secure. The Pros with this is it is fast, easy to deploy and not too expensive. The Cons would be that it is NOT widely available, you are limited to a single metropolitan area (not scalable).
Frame Relay - This is a real tried and true technology. You basically order a decent circuit at the Hub site, smaller circuits at each remote site, then configure a point-to-point interface between them and route across. Frame Relay has the concept of CIR or Commited Information Rate - this means how much bandwidth are you promised. For example, you might buy a T1 circuit, but only gaurantee the first 768K of it. The remainder of the of the T1 (the other 768K) is still there and useable - but they can't promise your traffic will make it. The Pro's - well documented technology, lots of resources for it. The Con's would be that it might be more expensive than either other option, and it is an older technology.
MPLS - this is a term I find scares people and it shouldn't. Really, you are buying a Layer 3 handoff - not an MPLS service. They give you a circuit (access charges will be similar to Frame-Relay) but you actually build a routed adjacency to the service provider's router (called the PE). Tehn you just route your traffic across. A very nice benefit of the L3 handoff is that it means you are fully meshed out of the box. The pro's are:full meshed network right away and Layer 3 routed handoff (easier to support for many). The downside is that it is at least as expensive as Frame Relay and depending on what routing protocols they support, the connection could get tricky initially.
Which is best for you I am not sure, you may want to do some thinking about what you can best support, and what the most important needs of your company are. I hope some of this helps.