Manual:Hotspot Introduction

From MikroTik Wiki
Revision as of 14:55, 7 October 2010 by Marisb (talk | contribs) (Getting an Address)
Jump to: navigation, search

Summary

HotSpot is a way to authorize users to access some network resources, but does not provide traffic encryption. To log in, users may use almost any web browser (either HTTP or HTTPS protocol), so they are not required to install additional software. The gateway is accounting the uptime and amount of traffic each client have used, and also can send this information to a RADIUS server. The HotSpot system may limit each particular user's bitrate, total amount of traffic, uptime and some other parameters mentioned further in this document.

The HotSpot system is targeted to provide authentication within a local network (for the local network users to access the Internet), but may as well be used to authorize access from outer networks to access local resources (like an authentication gateway for the outside world to access your network). It is possible to allow users to access some web pages without authentication using Walled Garden feature.


Getting an Address

First of all, a client have to get an IP address. It may be set on the client statically, or leased from a DHCP server. The DHCP server may provide ways of binding lent IP addresses to clients MAC addresses, if required. The HotSpot system does not care how client get an address before he/she gets to the HotSpot login page.

Moreover, HotSpot server may automatically and transparently change any IP address (yes, meaning really any IP address) of a client to a valid unused address from the selected IP pool. If a user is able to get his/her Internet connection working at their place, he/she will be able to get his/her connection working in the HotSpot network. This feature gives a possibility to provide a network access (for example, Internet access) to mobile clients that are not willing (or are disallowed, not qualified enough or otherwise unable) to change their networking settings. The users will not notice the translation (i.e., there will not be any changes in the users' config), but the router itself will see completely different (from what is actually set on each client) source IP addresses on packets sent from the clients (even the firewall mangle table will 'see' the translated addresses). This technique is called one-to-one NAT, but is also known as "Universal Client" as that is how it was called in the RouterOS version 2.8.

One-to-one NAT accepts any incoming address from a connected network interface and performs a network address translation so that data may be routed through standard IP networks. Clients may use any preconfigured addresses. If the one-to-one NAT feature is set to translate a client's address to a public IP address, then the client may even run a server or any other service that requires a public IP address. This NAT is changing source address of each packet just after it is received by the router (it is like source NAT that is performed early in the packet path, so that even firewall mangle table, which normally 'sees' received packets unaltered, can only 'see' the translated address).

Icon-note.png

Note: arp mode must be enabled on the interface where one-to-one NAT is used



Before the authentication

When enabling HotSpot on an interface, the system automatically sets up everything needed to show login page for all clients that are not logged in. This is done by adding dynamic destination NAT rules, which you can observe on a working HotSpot system. These rules are needed to redirect all HTTP and HTTPS requests from unauthorized users to the HotSpot authentication proxy. Other rules that are also inserted, will be described later in a special section of this manual.

In most common setup, opening any HTTP page will bring up the HotSpot servlet login page (which can be customized extensively, as described later on). As normal user behavior is to open web pages by their DNS names, a valid DNS configuration should be set up on the HotSpot gateway itself (it is possible to reconfigure the gateway so that it will not require local DNS configuration, but such a configuration is impractical and thus not recommended).


Walled Garden

You may wish not to require authorization for some services (for example to let clients access the web server of your company without registration), or even to require authorization only to a number of services (for example, for users to be allowed to access an internal file server or another restricted area). This can be done by setting up Walled Garden system.

When a not logged-in user requests a service allowed in the Walled Garden configuration, the HotSpot gateway does not intercept it, or in case of HTTP, simply redirects the request to the original destination. Other requests are redirected to the HotSpot servlet (login page infrastructure). When a user is logged in, there is no effect of this table on him/her.

Walled Garden for HTTP requests is using the embedded proxy server . This means that all the configured parameters of that proy server will also be effective for the WalledGarden clients (as well as for all clients that have transparent proxy enabled)


[ Top | Back to Content ]