Many interpret Portal authentication as a page that will be shot into a backstage agreement. They are two paragraphs of the same thing: the whole segment where users see and operate in browsers, and the entire section where access devices and authentication platforms answer. Distinguishing these two pieces, you know what to look at, and when receiving and inspection, what to check. The mixed result is that pages are not open and accounts are checked, and the account numbers are not checked and time is spent on the wrong direction.
The jump that the user saw.
Portal, which is responsible for drawing users to the authentication page, pushing them over the page, receiving information he fills on the page and making him connected to the Internet normally after certification has been passed. It is faced with people and browsers, so it is unclear to consider the speed of loading a page, end compatibility, failure tips. This paragraph is problematic because users have a direct feeling that the page cannot be pop up, the page is stuck or the filling is not done. It has nothing to do with the account number, the account numbers are all right, and the pages are not even available online.
The question behind the device.
The user of Radius does not see the end-delivery authentication request at all. The end-send is transferred from the access device to the authentication platform, which checks the identity and returns an acceptance or rejection result, while the end's strategy that should be obtained is taken back together with the equipment and the message, so it has to consider whether the request can be made, whether it responds quickly or if there are missing fields. In this case, users feel that the page can pop out and the information is filled up, but always point to a failure in authentication or a looping.
The parameters of the authorization are sent with the results.
The authentication platform does not just pass or fail. This terminal can use a bandwidth, which section of the network, how long it is valid and whether to make additional restrictions. These parameters are usually sent to access devices along with certification results, and they are carried out by the device on specific links. So the real location where the strategy works is on the side of the device, and the platform is simply telling the equipment how to do it. It is important to understand that: the platform looks at the strategy and does not work on the chain; there is a gap between the ability to execute the next release and the side of the device.
The billing is another line.
Authentication is an action, and the billing is a session. After authentication, users start to build a session on the counter, with periodic updates in the middle, ending when the user leaves or gets kicked off line. Online time and traffic are both coming out of this line. It is not the same thing as authentication shared identity and meeting identification. There is often a normal but uncosted certification in the project, or, conversely, because the respective configuration and status of each line is not right, rather than the authentication itself.
Two broken symptoms are different.
The failure of the Portal section, which is typically a case in point when the page cannot be shot out, the page can not be loaded and the user has failed to respond after submission, is usually unaffected because it is certified. Radius, which is typical of the page that is open normally and information that can be filled, but which fails or is over-timed after submission, may also affect users who are already online, depending on whether the side of the device is going back to the platform at every time. See which part of the event is first taken and then look down.
When checking, you have to look at the two types of records separately.
The key basis for judging the section of Radius is whether or not a request has reached the platform, at which point and where it is. The page pushes and clicks on this data reflect Portal. The record of authentication is first looked at when the access is required, because the request is the hardest fact that it can cut the problem directly into the network and device side or the platform side.
Both periods are to be written separately at the time of receipt and inspection.
The two paragraphs should have separate entries in the receiving and inspection list. Portal at least checks whether different terminals can eject pages, how long page load controls are available, and whether failure reminders are clear. Radius at least asks whether authentication requests will be stable, what level of success is, whether the release strategy actually works on the side of the device, and whether the forced downlines will be enforced. Writing them into a general authentication is valid, which would mean putting two completely different issues into a judgement and not locating responsibility when they arise.
You have to look separately.
The pressure source for the two paragraphs is different. Portal, which is closer to the front end and has been tightened when it reaches high, first page push and request reception; Radius, which is closer to the back end, first identification check and strategy test. When looking at a total amount of authentication, it does not see which one of these blocks is used to create a problem. So it is suggested that the volume of requests in both paragraphs be collected separately and separately over time, with either side of the line starting with a turning point, and the stretching and optimization move first. What size can be measured by actual hardware and punctures of the project, and the numbers on the data cannot be taken directly into account.
Some projects are just one step.
Not every project runs both of these. Some settings have access to a simple authentication capability, which requires the platform to perform an account check and the page is presented by the device itself; others just need to confirm that it does not go undefeated without a complex strategy. This simplification is correct, but it needs to know what has been simplified: often tactical flexibility and uniform management of particles. The initial project could be streamlined for quick entry, if the limitations resulting from this simplification are clearly written and the subsequent operations become complicated.
Putting the division of labour in the plan.
The last point is practical: the plan document should clearly draw the boundaries between these two paragraphs, and make clear which one is responsible and who is wrong. This is particularly true for multi-manufacturer projects, where access equipment is one, authentication platform is another, and each side of the problem is said to be the other. Writing the division and the decision in advance has a common basis for agreement when disputes arise, rather than relying on the word " executives ", on the spot.