SVWUX Wireless Network Topology & FAQ

Topology

The SVWUX 802.11b/g wireless network has several parts. One is comprised of a single Linksys WRT54G router located on a hilltop above Los Gatos. This router broadcasts an ESSID of svwux. Once a wireless node attaches to the ESSID, the router will provide basic DHCP information (everything except DNS and default gateway). This site could give coverage over most of the south and some of the west side of the valley. Unfortunately the omni-directional antenna used at this location collects too much noise and effectively blocks all but the strongest of incoming signals. If you have a 22db antenna you might be able to be 4 to 5 miles away and still connect. This site will need to be rebuilt with a number of high-gain antennas and a router before it will reach it's desired potential. This is in the planning stages.

Fundamentally, this means that all connections of any length must be done with high gain antennas (18db or more) on both ends. These antennas will have a correspondingly wide 1/2-power beam-width (18db = 20deg, 22db = 8deg, etc) which will give us some "wiggle-room" in antenna pointing and resiliency during earthquake shaking. This will also allow us to cover large areas of the valley floor from our hill-top sites.

Another part of the network is in the San Jose Emergency Operations Center (EOC) in downtown San Jose. We were in the process of connecting the San Jose EOC to the Los Gatos router, but this ran into difficulties due to an antenna conflict on the EOC roof. The San Jose EOC also has an omni-directional antenna to enable the Santa Clara County EOC to connect to the city EOC (a rather short hop away). This will also enable the San Jose Backup EOC to connect to the main EOC (across the parking lot). We believe we have solved the antenna conflict on the roof and can now continue our planned installations.

Our next goal is to establish a second hill-top router location, hopefully on the hills east of San Jose. This will give coverage over most of the east side of the valley. Once we get either a second hilltop router established or the Los Gatos repeater connected to the San Jose EOC we will be able to establish and Internet gateway for the entire network.

Once our Los Gatos router is rebuilt we can begin talking to a few other city's EOC's about connecting to our network.

One other important aspect of our network is that we will use IPv6 throughout, though not exclusively. We will use IPv4 alongside IPv6 but our IPv4 addresses will be non-routable currently. Our IPv6 addresses will be fully routable.

There are two wireless node types: client and gateway. A client is a node that does not have an independent Internet connection. A gateway is a node that has an independent Internet connection and wishes to share its bandwidth with other users connected to the SVWUX wireless network. (For more information about setting up a client or gateway, visit the "howto" pages for each.)

FAQ

The Part 15 v.s. Part 97 Question?

FCC Part 15 regulations govern low-powered non-licensed devices where Part 97 governs Amateur Radio. Part 15 is very limited in power where Part 97 is almost unlimited (relatively speaking). In 802.11b/g devices, channels 1 through 6 fall within the Amateur band and can be transitioned to Part 97 power levels by a licensed Amateur Radio operator. Unfortunately, Part 97 power levels bring with them a lot of restrictions on the kind of traffic that can be sent over that radio link. No encryption can be used (an exception is made for controlling satellites, which does not help us any). WEP encryption is allowed if the key is published in a public place, say in a web page. The problem with this is that WEP is so weak as to be useless. The real nail-in-the-coffin for the inability to use encryption is that no hospital, with current HIPPA data security requirements, will do any serious communications across an unsecured link. This goes for corporations as well.

Another issue is that during a true emergency the rules may be suspended to facilitate communications, so all power levels can be increased. This is true, but of limited help. We want to avoid having gear in our network that we only turn on during a disaster because if we only turn it on during a disaster we won't know it's broken until it's too late to fix it. We would prefer to run our network the same way at all times. One aspect of this that we are considering is that some gear can change it's power level through a software setting.

Do we need access to the best repeater sites?

Most large and popular repeater sites are very well located, but they are also very crowded and very noisy. We would prefer more non-traditional sites where noise is not a big problem. We think we will be able to make up for shortcomings in sight-lines so we will take that trade-off.

Will our network be operational only during disasters?

No, it will be operating at all times. If we only operate it during a disaster then we will only find out it's broken during a disaster. Then it's too late to fix it.

We want our network to operate the same way that Amateur Radio repeaters work; they are available for casual use until an emergency, when they are reserved for emergency use.

How will we keep our network from being overloaded?

We will define a number of SSID's (network names) that are available everywhere on our network. Each SSID will be associated with a different VLAN, each with different QoS^? and Traffic Shaping rules (priorities). For example, First Responders will have all of the bandwidth they need, NGO's will have the next priority, corporations the next, and the general public the last.

Will this network be for governments only?

No. We feel strongly that disasters effect all of society and therefore all of society should be involved (at some level) in response and recovery. A major part of Silicon Valley's continued success is based on the economic resiliency of it's corporations (not to mention my retirement fund, my kid's college education funds, etc. :-). Therefore we want to provide a mechanism to help this cooperation.

How will we connect to the Internet?

We can provide Internet access from any of a number of locations within and from outside Silicon Valley. Any interconnection between the EWN and the wider Internet will be able to provide Internet access to the entire EWN. Our goal is to have a number of interconnect points both inside and outside the Silicon Valley.

How will we provide redundancy and reliability in the EWN?

This is a difficult question. The common, knee-jerk reaction is to have no single-point of failure. But the complexity involved in providing a failover mechanism at every point can often be very fragile, resulting in a net loss of reliability. Our contrasting view leaves every network link as simple as possible but provides redundancy at a higher level, namely at the level of Internet Routing over many connections each utilizing different technologies.
to top