Deliver Applications Faster With Automation

Introduction

Over the past five years, application delivery networking has evolved drastically. Gone are the days when a simple DNS resolution-based mechanism was considered to be load balancing. Application delivery controllers (ADCs) have become an intelligent component of the network and have changed the way applications are provisioned.

In the past, applications were developed and deployed at a moderate rate. But with the advent of DevOps, application teams are getting more aggressive and are rolling out applications at an unprecedented volume and speed. The biggest challenge is to make these applications production-ready immediately after they have been developed. New applications in an enterprise-grade environment need to be able to scale to handle traffic from hundreds or thousands of users, and they demand high availability. Enterprises leverage ADCs to improve the reliability, security, and performance of applications. Essentially, how quickly an enterprise can make an application production-ready depends on how quickly it can configure ADC for it.

Enterprises are also striving hard to reduce time to market. Deploying an application requires configuring a virtual IP (VIP) for load balancing and creating DNS records for name resolution. Application teams generally have to submit multiple requests to deliver an application. The current approach of allocating a VIP for an application is quite isolated. Different IT operations team work in their own silos, making the process inefficient and delaying delivery.

For enterprises to overcome these challenges, they need a new approach to managing their application delivery infrastructure. By reducing reliance on discrete, manual processes and employing an advanced ADC management, automation, and orchestration solution, enterprises can simplify and accelerate application delivery to keep up with rapidly growing business demands.

The bottleneck: ADC deployment impacts application delivery

It is easy to say that application deployment should be faster. But it is another thing to realize that with the traditional approach, it is impossible to speed up delivery. When an application team has to deploy a new application, or modify existing applications, the network team receives change requests at several layers (ADCs, DDI, firewalls, etc.). It usually takes multiple weeks to provision an application. Recent research by Enterprise Management Associates (EMA) reports that 47 percent of organizations take more than a week to deliver an IT service request, and 29 percent require more than two weeks.1 There are many challenges that delay application delivery. A two-week delay to create a VIP is no longer acceptable. The data center—and specifically ADC deployments—must be agile enough to meet the new expectation levels that the business demands.

Limited collaboration of application and network teams

Enterprises manage new IT requests or changes using methods such as ITIL. They use ITSM tools, such as BMC Remedy, ServiceNow, and external ticketing systems, to manage changes for the entire infrastructure stack. Application teams have to submit requests to multiple teams to deploy an application. At the same time, network engineering teams are often piled up with many other incident-management issues. This approach is somewhat similar to a car parked in a narrow lane, where the second car can only move if the first car moves. Application and network teams work in silos, which delays service delivery and results in revenue loss to an enterprise.

Lack of automation and orchestration

In a typical application-provisioning scenario, when an application is rolled out, the application team submits a request for a VIP. Generally, the request contains details such as application name, VIP, port number, and load balancing algorithm. This request is then pushed to the network team. The network team verifies the information, and if any information is found to be missing or wrong, the request is pushed back for rectification. The network team then refers to an IPAM solution to get a free IP, and the selected IP is provided back to the application team. In the worst-case scenario, if the application needs to be rolled back again, a new ticket gets started. This complete process is iterative and time-consuming. It is a manual approach that is labor-intensive and involves repetitive tasks.

Error-prone manual deployment

All the information entered, including the application name, VIP, port number, and so on, is done so manually. There is no way to validate that the application is functioning once a VIP is provisioned. What if an application team wants to make changes to an existing application? How can they ensure that the new deployment has met organizational standards? These are some critical questions that remain unanswered. Humans make mistakes—there is no question about that. Research by Dimension Data shows that the number one root cause of network outages is human error, with 37 percent of outages resulting from configuration or other human errors.2 A survey of mid-to-large companies by Avaya found that 82 percent of respondents experienced some type of network downtime due to human error.3

1 EMA, “Network Management Megatrends 2016,” April 2016
2 Dimension Data, Network Barometer Report 2016 (Accessed January 23, 2017)
3 Avaya, Network Downtime Results in Job, Revenue Loss, March 5, 2014

Automation Is Key to Accelerating Application Delivery

The limitations of manual management are clear. As the business need for high-performance applications continues to grow, enterprises that rely on traditional, manual processes will only fall farther behind in their ability to deploy ADCs quickly, limit errors and any resulting outages, reduce time to market, and minimize the overall management burden on IT resources.

Because application deployment in enterprises is complex, time-consuming, and cumbersome, one of the top priorities in speeding up the process must be to bridge the isolation across the multiple teams that are involved in application deployment. Next, automation should be introduced for creating virtual IPs. Automation becomes critical when multiple stakeholders have to collaborate effectively to provision an application.

These goals can be accomplished by embracing a new approach to ADC management-one that replaces manual processes with automation and orchestration. With the right ADC management and automation tool, enterprises can use granular access control, self-servicing capabilities, and standardized workflows to manage various application deployment scenarios for brownfield and greenfield environments with much greater speed and efficiency.

A Complete Solution for ADC Deployments in Brownfield and Greenfield Environments

AppViewX’s Application Delivery Automation Solution is a complete ADC management solution that leverages best-in-class automation and orchestration methodologies to enable faster application deployments in brownfield and greenfield environments. As part of the AppViewX’s Application Delivery Automation Solution, the AppViewX Application Provisioning System (APS) provides a unique, self-service, and form-based approach that brings network teams and application teams to the same table.

APS facilitates collaboration among cross-functional teams with a multi-stage workflow configuration system and ensures rapid application deployment and post-deployment validation. It enables self-service and automates the creation and modification of VIPs.

Granular, role-based access control

Role-based access control (RBAC) provides granular, object-level access and enables the enterprise to define specific roles for performing actions across the network infrastructure. Network engineers can design and develop a self-service form for making default changes and creating or modifying new VIPs. These simple forms can be delegated to different application teams for quick and easy configuration of VIPs for various application deployment scenarios.

Simple self-service forms

Another priority is ease of use. Using APS, simple forms can be designed and delegated to different application teams. These forms are highly customizable and can be hard-coded based on organizations’ requirements. For instance, if an enterprise wants its users to use only one ADC for application deployment, the form can be hard-coded to make only that ADC visible to the user. Similarly, fields such as name and load balancing algorithm can be hard-coded.

APS does not require a steep learning curve. Designing a form is also very easy, and a simple form can be designed in minutes. APS comes with pre-defined worker scripts to be associated with the form fields.

Once an application is rolled out, the application owner can fill in details about the application and submit it. After the form is submitted, it talks to the relevant ITSM system via RESTful APIs to create a ticket that contains all the information the application owner entered. This ticket is then pushed to the network team, who verifies the information and approves it if it is satisfactory.

APS receives the approval via the ITSM, talks to the IPAM system via RESTful APIs, and picks up a free IP. Not only is an IP is fetched for the VIP, but DNS records (A record and PTR record) are also created for the application in the corresponding IPAM system.

The network team can then review it if required. Once the final approval is received, the configuration file is built based on the business logic and pushed to the ADC to make the necessary changes. Multiple levels of approval can also be configured based on enterprise policies. A scheduler-based option is available that enables the configuration to be pushed on a particular date and time. Alerts can be configured to trigger an email or SMTP trap for notification and monitoring purposes.

Rollback configuration

APS provides the flexibility to undo ADC changes when necessary, ensuring that the ADC returns to the same state it was in prior. This is possible because it takes a snapshot of the configuration file before implementation and commits it if rollback is required.

Post-validation

The solution empowers teams to validate whether applications are performing as expected. Everything from simple commands like ”curl” to complex Python scripts can be defined. Post-validation serves as a sanity test for the applications to make sure they are delivering what they are expected to. It can also be used as a troubleshooting mechanism when applications are not working.

Application topology view

The application topology view gives application and network teams visibility into the application infrastructure so they can troubleshoot application-related issues faster. It provides a network map of the ADC infrastructure with its complete hierarchy. Users can recursively look up the pool members (end servers) where multiple devices are handling traffic for one application, and they can take various actions, such as performing a backup, restoring a configuration, viewing a configuration, and viewing change and audit logs.

Conclusion

It is easy to say that application provisioning should be faster, but huge interdependencies across different teams make it is impossible to speed things up using a traditional approach. As a result, application deployment is delayed-often by as much as 7-10 days-creating an unacceptable situation.

AppViewX APS has helped many customers to automate and orchestrate application delivery tasks. A simple, self-service form-based approach has assisted numerous customers in shortening their application deployment window from weeks to days or even hours. AppViewX provides the ability to define workflows for approval, validation, and implementation to help companies deliver applications faster.

Technology integrations

AppViewX integrates with leading technology providers to provide state-of-the-art automation and orchestration capabilities.

Technology type Vendors
ADC A10 Networks, Akamai, AVI Networks, Brocade, Citrix, F5 Networks, Radware
ITSM BMC Remedy, HP Enterprise, ServiceNow
DDI BIND, Bluecat, InfoBlox, VitalQIP

Learn More

To get started with AppViewX, start a free trial. AppViewX supplies prepackaged templates to address various application deployment scenarios, including the creation or modification of virtual IPs, to help businesses provision applications faster. Some standard templates are available in the AppViewX GitHub community (note that you must have an AppViewX instance in your environment to try the templates). Contact AppViewX for more information.

Tags

  • Application Automation
  • NetOps Automation
  • Network Automation