The bulk dialer allows you to make calls out to a large set of numbers, and any that are successful are forwarded to whatever destination you choose. This article is going to go over the steps I suggest to take in order to a Cluster Dialer setup when using several methods. We, Dialerking Technology do things a little differently when it comes to this to prepare for the worst case scenario that a server may have problems and have to be taken offline. The numbers to call are grouped into “campaigns”, which are sets of numbers with the same settings.
The clustering feature provides a mechanism for seamlessly distributing call processing across the infrastructure of a converged IP network. Clustering facilitates redundancy, provides transparent sharing of resources and features, and enables system scalability.
Welcome to Dialerking Technology’s guide on setting up a Cluster Dialer, optimizing your call center operations with Load Balancing, High Availability, and more. This guide covers essential aspects of deploying and configuring a Cluster Dialer system for improved efficiency and reliability.
What Are Clusters?
A cluster comprises a set of servers that share the same database and resources. You can configure the servers in a cluster in various ways to perform the following functions:
- Database publisher server
- TFTP server
- Application software server
- Primary call-processing server
- Backup call-processing server
When you install the dialer software on servers, you specify which servers belong to the same cluster. Using the Service Activation window, you specify which server performs which function for the cluster. You can dedicate a particular server to one function or combine several functions on one server, depending on the size of your system and the level of redundancy that you want.
A minimum starting cluster would consist of a single Telephony server and a combined Database and Web server. A three server cluster, whereby there is at least a single server dedicated to performing the role of a Database, Web, and Telephony server is the more robust option and requires little to no
reconfiguration as the cluster grows in size
Each cluster can have only one database publisher and usually one TFTP server (either separate or combined). Other servers in the cluster that subscribe to the publisher database maintain their own local copies of it.
Understanding Cluster Dialer Setup
A Cluster Dialer setup involves creating a robust infrastructure that distributes call loads across multiple nodes. This ensures high availability of agents as well as scalability.
Cluster dialers can achieve connection rates ranging from 10% to 30% of attempted calls.
ViciBox “Express” System
A ViciBox “Express” system consists of all major roles including Database, Web, and Telephony server. Using the minimum hardware specifications listed above, this would be well suited for a call center with up to 15 agents doing outbound calling with call recordings at a 4:1 ratio. When used for
inbound the same minimum hardware specifications are acceptable for up to 25 agents.
Please note that the figures presented are a conservative guideline and that your actual system capacity is heavily determined by the number of agents, number of leads, concurrent call count, dialing ratio if applicable, and the general capacity of the hardware. It is for this reason that additional specifications were provided should you run into any minor load issues. Particular attention should be paid to the size of
the hard-drive(s) installed in the system if you record your calls. The most frequent support issue we find from new customers is a crashed system from the hard-drive being full of recordings.
ViciBox “Cluster”
Clustering is where multiple individual servers perform specific roles within ViciDial. Those roles are Database, Web, Telephony, and Archive server. Clustering should generally be used for an install consisting of more than 20 outbound agents or 30 inbound agents. If you do not know or are unsure of
whether you should perform a clustered install please consult with the free community-based ViciDial forums or paid-for ViciDial Group for support. Please be prepared to explain your planned installation
including number of agents, call type (Outbound/Blended/Inbound), expected dialing ratio or concurrent call volume, archival requirements, and any other unique processes or requirements you may have. This will help generate a more accurate answer to your questions.
Load Balancing for Optimal Performance
Load balancing is a critical component of a cluster dialer setup. It is designed to distribute incoming call traffic across multiple nodes efficiently. This ensures optimal performance by preventing any single node from becoming overloaded, thus maintaining high availability and responsiveness in your call center operations.
Load balancing refers to the process of evenly distributing incoming call traffic across multiple servers or nodes within a cluster. The primary goal is to optimize resource utilization, maximize throughput, minimize response time, and avoid overload on any specific node. This is particularly crucial in call center environments where call volumes can fluctuate throughout the day.
Implementation of Load Balancing
Implementing load balancing requires specific strategies and tools to effectively manage call distribution. Here are key considerations for implementing load balancing in your cluster dialer setup:
Load Balancer Selection
Choose a reliable load balancing tool or software that suits your infrastructure and requirements. Popular choices include HAProxy, Nginx, and software-defined networking (SDN) solutions.
Traffic Distribution Algorithms
Load balancers use various algorithms to distribute traffic, such as Round Robin, Least Connections, or Weighted Round Robin. Select an algorithm based on your specific workload characteristics.
Health Monitoring
Configure health checks to monitor the status of each node. The load balancer should automatically route traffic away from unhealthy nodes to maintain system reliability.
Session Persistence
For applications requiring session persistence (e.g., maintaining user sessions), configure the load balancer to ensure that subsequent requests from the same user are directed to the same node.
Scalability Planning
Design load balancing architecture with scalability in mind. Ensure that the load balancer itself is scalable and can handle increasing traffic volumes as your call center grows.
Best Practices for Load Balancing
- Regularly monitor and analyze traffic patterns to optimize load balancing algorithms.
- Implement failover mechanisms to redirect traffic in case of load balancer failure.
- Consider geographic load balancing for multi-region deployments to minimize latency and improve user experience.
Ensuring High Availability
High Availability refers to the ability of a system to remain operational and accessible for a high percentage of time. It is typically measured as a percentage of uptime versus downtime. High Availability (HA) is a critical aspect of any Cluster Dialer setup, ensuring that your call center operations remain operational with minimal downtime. In this section, we delve into the details of what high availability entails and how to implement it effectively. For mission-critical applications like call centers, achieving high availability is paramount to maintain service continuity and customer satisfaction.
Components of High Availability
Achieving high availability involves implementing redundant components and failover mechanisms:
- Redundant Servers
Deploying multiple server instances that mirror each other’s functionality.
- Network Redundancy
Utilizing multiple network paths and connections to ensure continuous connectivity.
- Load Balancing
Distributing call traffic across redundant servers to prevent overload and single points of failure.
- Failover Mechanisms
Automatically switching to backup servers or systems when primary systems fail.
Cluster Dialer Setup By DialerKing Technology
ViciDialCluster Dialer Setup ViciDial scales horizontally for all roles except the Database. For this reason, you will need to spend extra time deciding on the size of your Database server. Easily expand your call center capacity by adding new nodes to the cluster, accommodating growing call volumes.
The specifications we recommend for a database server up to 150 agents is as follows :-
•Quad-Core CPU 2.0Ghz
•8GB ECC Ram
•4-drive 300GB+ 15K SAS RAID-10 on a hardware battery-backed caching controller (LSI9260, etc)
•Redundant Power Supplies on a battery back-up
To add to the above specification for handling up to 300 agents would be as follows:
•Two Quad-Core CPU 2.0Ghz
•16GB ECC Ram
For a 500 agent Database server, adding to the specification, would be as follows :-
•Two Octa-Core CPU 2.0Ghz
•32GB+ ECC Ram
•4-Drive 240GB+ RAID-10 on SSD with SSD-optimized RAID card or software Linux RAID
•Two Seperate 500GB enterprise-grade SATA drive for the OS in a RAID-1
Clustered dialer setups are adaptable to changing business needs, allowing for seamless integration of new technologies and features. Contact DialerKing Technology for expert guidance on designing and deploying cluster Dialer Setup for your call center infrastructure. A stable and responsive smart dialer solution system translates to better customer interactions and satisfaction.
