Large Scale Applications Deployment Considerations

Autonomy in Total Cost of Ownership (TCO) Management: Departments seek control over the total cost of ownership associated with their assets and applications. This involves understanding and managing costs throughout the entire life cycle, including acquisition, operation, maintenance, and disposal.

Decentralized Management: The organizational structure is decentralized, allowing each department to independently handle its resources. This decentralization can lead to greater flexibility and responsiveness to specific departmental needs.

Life Cycle Management Autonomy: in large scale deployment, it is typical to have multiple departments, each responsible for managing the entire life cycle of its assets and applications. This includes planning, development, deployment, maintenance, and eventual retirement or replacement. They need complete independence from other department and cannot rely on any central authority for that to provide fast response time to their users.

Flexibility in Bandwidth Management: he applications require the flexibility to scale their bandwidth as needed without having to rely on the ADC vendor to license every adjustment. This implies a self-service model where the organization or application owners can dynamically allocate and modify bandwidth resources based on demand.

Autonomous ADC Instance Management: Similar to bandwidth, the organization seeks autonomy in managing ADC instances. This could involve the ability to provision, scale, or de-provision ADC instances without direct intervention from the ADC vendor. This autonomy allows for quicker response to changing application requirements.

Scalability Requirements: Applications at large scale often have varying requirements for bandwidth and ADC instances. With thousands of applications to manage, some applications may need highly dynamic bandwidth and numerous ADC instances. “Highly dynamic” means that most of the time, the application has a stable average, but at times, the consumption spikes considerably – 10x or even 100x. Emergency applications are good examples of such cases. It does not make sense to have a fixed bandwidth allocation for such applications, as it is a waste of resources. The system should accommodate these diverse scalability needs.

Integration with Cloud and Hybrid Environments: Considering the mention of large-scale applications, it is important to consider whether these applications are deployed in multiple public and private clouds. The management system should seamlessly integrate with these environments, providing a unified approach to bandwidth and ADC instance management.

Cost Management and Optimization: With autonomy in bandwidth and ADC instance management comes the responsibility of cost control. Organizations need mechanisms to monitor, analyze, and optimize costs associated with these resources to ensure efficient utilization.

Automation for Rapid Adjustments: Automation is essential for rapid adjustments in bandwidth and ADC instances. Automated processes enable quick scaling or reduction based on predefined criteria, helping to prevent service disruptions, and ensuring a responsive infrastructure.

Granular Bandwidth Allocation: To efficiently utilize resources, there is a need for granular bandwidth allocation. Applications should have the capability to allocate bandwidth with high granularity, adjusting it based on specific requirements without being constrained by fixed increments.

Self-Service Model: Implementing a self-service model enables application owners or administrators to manage bandwidth and ADC instances through an intuitive interface or automation tools. This reduces the dependency on the ADC vendor for routine operational tasks.

Monitoring and Reporting: Implementing robust monitoring and reporting capabilities is crucial. This allows organizations to track the performance of bandwidth and ADC instances, make informed decisions, and optimize resources based on historical data.

1. Global Elastic Licensing (GEL): With Radware’s GEL model, organizations can purchase one single license of ADC capacity, and dynamically break it down to as many ADC instances as required, allocate capacity per instance – as required, scale up or down that capacity, and even decommission ADC instances, using their capacity in hybrid environments – as required.

2. Cyber Controller as single pane of glass for management, monitoring & analytics:

• Granular RBAC
• Proactive actionable monitoring and analytics
• Single pane of glass for all ADC instances and application services
• Clear breakdown between App-Network-Client Round Trip Time (RTT)
• Drill down to Root Cause Analysis
• Supporting Multi Cloud

3. Template Engine: templates allow deploying a predefined configuration of instances and applications instead of manual configuration for:

• Simple Deployment & Scale
• Reduced configuration errors
• Best practices sharing

4. Advanced GSLB and supporting High Availability mode for optimal failover performance.

5. Alteon auto-scaling cluster enhances the scalability and efficiency of the ADC service and allows handling the following type of scenarios:

• Gradual growth – An ADC cluster can start with a couple of Alteon devices that meet the initial traffic needs. As the application traffic increases, the scaling cluster handles it by automatically adding more Alteon instances to meet the requirement.
• Planned or un-planned traffic surge – When the ADC cluster is suddenly presented with a significant increase in traffic that exceeds the current cluster capacity, additional Alteon instances are temporarily provisioned to handle that increase. Once the traffic load goes back down, the additional instances are automatically decommissioned.

6. Alteon natively support most common environments such as: VMware (vRa/vRo), OpenStack (Heat) and Ansible Automation & Orchestration tools.

7. Public Cloud and multi cloud support: With Radware unique Global Elastic licensing we are supporting the most common public clouds such as Azure, AWS and GCP using the same Instance image among all types of deployment without compromising on the feature set or the management life cycle.