A practice of converting a system deployment from a series to a parallel system

Chapter 1 Status Quo Analysis

The deployment of a telecom Intelligent Network IP (smart peripherals) platform is as follows:

There are two sets of IP addresses: IP1 and ip2.

The IP hardware configuration is three Excel and four ibm pc servers. Two signaling points are configured and two sets of IP addresses are enabled. The process allocation is as follows:

1. the ISUP, vn1, and vn2 processes of IP1 run on the IBM x346 host, and the lifekeeper is responsible for the dual-host;

2. The Cn1/cn2/tcap process of IP1 runs on the IBM x3650 machine, and the lifekeeper is responsible for dual machines;

3. the ISUP and vn1 processes of ip2 run on the IBM x346 machine, and the lifekeeper is responsible for dual machines;

4. The Cn1/cn2/tcap process of ip2 runs on the IBM x3650 machine, and the lifekeeper is responsible for dual machines.

The IP process connection box is shown below (the figure above is IP1, and the figure below is ip2 ):

The process port distribution is shown in brackets:

1.1 Network physical connection Diagram

The two ibm3650 servers are deployed with ipcn and TCAP, and are mutually active and standby. The two IBM x346 instances are deployed with isupgateway and Vn, and are mutually active and standby.

1.2 logical diagram of the network connection Function

 

IP1-ISUP and IP2-ISUP processes run on IBM x346 machines, master and slave, IP1-CN and IP2-CN run on IBM x3650 machines, master and slave, such:

The IP1-VN process is installed under the ipcn user of IBM x346 machine, the IP2-VN process is installed under ipcn2 user of IBM x346 machine, and the IP1-CN process is installed under ipcn1 user of IBM x3650 machine, the IP2-CN process is installed under ipcn2 user on the IBM x3650 machine.

1.3 Risk Analysis

1. The running status of the IP1 platform and the ip2 platform is in connection. If a problem occurs on one platform, the running status of the other platform is affected. This reduces system reliability.

2. ip_cn uses a floating IP address to establish a socket connection with ip_isup. If the floating IP address fails, ipcn cannot connect to ip_isup correctly, which will cause normal sound releasing.

3. When lifekeeper fails to determine the network conditions, it starts the master and slave processes, which causes the master and slave ip_vn to compete for the Excel resources, resulting in abnormal sound releasing.

Chapter 4 adjustment suggestions

Divide IP1 and ip2 physically into independent systems and run them in parallel. Disable the socket connection between ip_cn and ip_vn through a floating IP address, and use the real IP address instead. When ip_cn and ip_vn are started, the cluster mode is used to prevent the master and slave processes from starting at the same time.

2.1 network physical connection diagram 2.2 network connection function logic diagram

Ibmx346 is used as the slave of ibmx3650.