The deployment of a WiFi certification system is not particularly complex from a technical point of view.——Select a certification software, with network devices and access to Portal pages. Tests can be done several rounds. But from the project management point of view, pits are much more numerous than they look. Because WiFi certification is not an independent IT system that cuts across network layers, application layers, front-end display levels, involves network equipment suppliers, authentication software providers, engineering implementers, operators ' multiple roles, and any disconnection of links may lead to project extensions or back-to-line work.
: Three things to look at outside the list of functions.
Most candidate typesWiFi certification systemIt starts with the functional list: which certifications are supported, whether there is an audit log, if you can access RADIUS and if you don't support multilevel management. That's right, but not enough. The functional list only tells you about this system."What can I do?"I can't tell you."How much trouble is it?"and"What if something goes wrong?"。
Three things should be added to the selection. One is the complexity of configuration and management——Not the more functional systems are, but the easier you can configure with the functionality that you have. Some systems have a lot of functionalities, but a simple authentication strategy needs to be changed several times between front and back desk, saving them for refreshing and having very poor transport experience. It is suggested that manufacturers should be asked to provide real management backstages for you to operate, not just to see the demonstration video during the selection phase.
Second is the ability to diagnose common failures. The system has a problem sooner or later, and the operator can quickly locate it. Is the log detailed? Are there visual monitoring panels? Are there built-in detection tools for malfunctions (such as network connectivity tests, RADIUS service accessibility tests, database connection tests) that are not normally needed, but when you get a call at 2:00 in the morning saying WiFi certification is dead, they save up a lot of time for checking.
The third is the manufacturer's technical support capacity and responsiveness.——The technical support team of the manufacturer will be able to provide a professional response. If the technical support of the manufacturer is not kept up during the selection, the level of support will not be better after formal collaboration.
Implementation phase: Programme evaluation is more important than programme writing.
Many projects are implemented by the manufacturer who wrote them for A, and he looks at it."Sure."The risk of this process is that the scheme written by the manufacturer is based on their standard template and does not take full account of the particular circumstances of the site. A person who approves may not have enough knowledge of the technical details of WiFi certification to see potential problems in the program.
The correct approach is that the programme evaluation must be attended by network engineers from the first party, IT transport manager, and information security chief. The vendor does not send sales engineer but implementation engineer to confirm each technical detail of the program in a item-by-item manner if it matches the situation on the ground."Programme adoption"Four words, but a summary with clear lines of action——Who needs to provide what information before what time, who is responsible for identifying which configurations, each with a clear head and deadline.
There is also an important management action in the process: regular (at least weekly) alignment of progress and problem synchronization.——The sooner the small problem becomes visible and resolved, the less impact the project as a whole. Project managers should be concerned about what is currently blocked, whose work has not been done, and what resources are needed to drive it. Instead of looking at the percentage of completion on the plan sheet.
Test phase: do not just perform functional tests.
Most projects have been tested to stop"Could not close temporary folder: %s"——Whether authentication can be completed, whether Portal pages can be displayed properly and the billing is correct. But WiFi certification system tests are much more than these. At least a few dimensions should also be covered: an anomaly scenario test (a network break during authentication process, text message code overtime, and system performance when issuing certification to reach maximum), equipment compatibility test (at least the type of device that covers TOP 20), security test (the Portal page contains XSS holes, authentication interfaces have SQL infusion risk, HTTPS configuration compliant), performance test (mock peak period and pressure test for sending authentication).
The test participants are not just the IT department.——A few representative real users should be asked to do the user acceptance test, observe their actual behaviour in the authentication process, and record where they are stuck and what is confusing. These feedbacks from real users often reveal problems that the testing team does not notice.
Up-to-line: Preparation for rollback programmes is more important than preparation for back-up programmes.
Everyone will write the online program.——The time is a few days to cut the flow, how long to observe and when it's clear that the problem is solved, but the rollback program is often very general:"If you have a serious problem, roll back to the old system."How do you roll back? How long does it take to roll back? Will you lose the user-certification data for this period after rolling back? These questions will be very confusing if you don't think back online.
A qualified rollback option should at least include: trigger conditions for rolling back (e.g., certification success of less than 95%, average certification time of more than 10 seconds, system resource use of more than 90% lasting 5 minutes), specific operational steps to roll back (dNS cutting back old IPs or gateways to old configurations), expected time-consuming per step, operator and contact information, and standards for checking back rollers. The scheme should be printed out by the operator rather than in a document on the computer desktop.
We should have one on the line."Stabilization observation period"During this period, the project team is on duty 24 hours a day and is ready to deal with sudden-onset problems. Every problem that comes up during an observation period, no matter how big, should be recorded in a list of issues, identifying those responsible and tracking it down. After the observation period, the project will be officially operational and the daily crews take over.
Delivery management: documents have longer life cycles than codes.
After the project is finished, it's not just a system, but a document. But in actual projects, files are often the last to rush out, and they are of poor quality, even if they are changed to a title by the standard manufacturer's documents. This leaves a huge risk for later transmissions and troubleshooting.——The recipients did not understand how the system was deployed, what the configuration parameters meant and where to start checking when problems arose.
At least these documents should be delivered: system deployment architecture (indicating the IP, role, port, dependency of each server), detailed configuration instructions (means and recommended values for each configuration parameter), a guidebook on how to operate a transport function (daily check, common trouble-processing process, back-up recovery operating steps), interface files (back-to-back external system interface specifications and call examples). These documents are not just thrown out once they have been written and should be continuously updated during system feeds.
WiFi certification systemThe project is managed in a cross-sectoral, multi-supplier and inter-technical process. Technical programmes are written by anyone, but whether the project can be lined with expected quality within the given time frame, within budget, by managing performance and risk control awareness at every step of the project process.