Part V in a series on how to deploy a city-wide wireless WiMAX network
All the work put forth to plan, design and install a quality WiMAX network can be for naught if the network is not proactively monitored and properly maintained.
Cities and service operators are likely to use WiMAX networks as a delivery method for critical data and voice services (such as public safety) and having a network actively monitored and maintained is critical to its proper performance and overall success.
In a large metro wireless network, there will invariably be dozens, if not hundreds of devices to monitor. These devices can include routers, switches, access points, CSU/DSUs, point-to-point links and other related items. The ability to manage remotely and monitor remotely these devices varies from one manufacturer to another. For the most part, today's leading vendors' infrastructure allows for remote SNMP management. The ability to get "into" the network and actively monitor various items is paramount to keeping the system operational to a carrier class level.
Ideally, the operator or integrator developing the network will install a remote server at the customer premise loaded with a hardware client VPN. This local server allows the end customer (albeit a city, school, service operator or an enterprise) to be monitored remotely.
Once the server/VPN is installed, a remote monitoring version of SNMPc (or something similar) can be installed. This solution allows the network elements to be polled locally as opposed to polling individual elements via the Internet, which would be constant polling. As the remote software polls the network, it sends a summary over the Internet to the monitoring firm, which interfaces with their main monitoring software.
Monitoring Set-up
With the first level of monitoring, the monitoring firm can track whether a device is responding or not responding. This basic monitoring advises if the network elements being monitored are operational or not.
Secondly, the baseline RSSI levels for each radio should be set in order to establish baselines of performance using RSL (receive signal levels). These levels should be recorded when the radios were installed. A good integrator will know what realistic or non-realistic receive levels are for particular radios.
Next threshold traps are installed within a 25% range of peak performance and that level is set as the trigger to create an alarm. This range of 25% allows the radio to continue to perform, but advises of a problem before the radio link goes down.
When a trap has been triggered it typically creates a graphical impression at the monitoring firm's NOC as well as sets off alarms to critical personnel about the alarm. The out-dialing capability is common and allows personnel to be notified 24/7.
A good integrator will also set up network collision tracking and monitor all collisions of data and loss of data in the network itself. Also the monitoring firm should monitor switch ports the radios are connected to in order to determine if there are problems at that level.
The integrator or firm that provides the monitoring service should also provide daily, weekly, monthly views of network performance including packet loss, RF quality so the network owner has a view of network activity, quality and overall performance.
Network Maintenance
Depending on the alarm error some problems can be repaired remotely. Recycling the radio and refreshing profiles or settings can sometimes fix a problem. There are however, field activities relating to maintenance that need to be planned for and undertaken. Such items include:
- Proactive updating and testing of the latest firmware
- Semi-annual inspection of all network components (both indoor and outdoor) to insure that weather seals, cables, ice breaks, electrical, UPS and other main network components are operational and not at risk.
- A local maintenance capability should also be available to be dispatched when problems do occur that cannot be fixed remotely.
- The local support personnel must be capable of connecting into the equipment via serial or Ethernet connection and resolving local issues.
- The personnel also has to have all necessary information to gain access to the site during off hours as well as have access to tower riggers in case of a lightning strike or antenna failure.
A service level agreement (SLA) should be developed by the city or operator to manage the commitments of the monitoring and maintenance company. It is not uncommon for service to be required within 2, 4, 6 or 8 hours depending on the critical nature of the network.
