Hybrid Devops – Software Defined Disruption
Businesses have been and will continue to be disrupted by software agility and innovation. If you have any questions, just ask, if they are still in business, Movie Rental Companies (Netflix), Taxi Companies (Uber), and Retail Companies (Amazon) to just name a few areas (companies that disrupted an industry with Software). Software defined disruption has changed the landscape and continues to drive tremendous business value like never before. What’s most exciting is that we have not seen anything yet compared to what the Internet of Everything (IoE) will disrupt! To understand software disruption better and determine the innovation opportunities it helps to take a look at the typical devops model today, challenges, and opportunities.
The typical devops model is represented the figure below:
The main concerns with how devops works today is that each application development team has to build and integrate each component and most enterprises have bought a solution that is less than ideal across the entire lifecycle but addresses 2 or 3 areas well. Each developer has to leverage the project management solution of their enterprise, along with developer tools and source control. The IDE of choice typically has a custom integration required with each set of tools and software models deployed in the enterprise. Continuous Integration and Deployment software and frameworks are typically not integrated with the existing SDLC tools and rarely take into account systems outside the enterprise data center or the internal systems dependencies. This is typically left to the developer as a problem to solve. Application orchestration and service assurance are difficult for developers to integrate into or support directly from their SDLC so they have to create this capability or determine another method or external service to enable this capability. Issue management is typically the “IT Help Desk” tool and not integrated into the application services. New devops capabilities require this flow to be automated and integrated. In addition these development needs, collaboration is becoming more and more critical to the software lifecycle and devops tools are integrating collaboration into each aspect of the devops lifecycle.
In addition to these devops short comings, the devops lifecycle actually needs a boast. Today, it still takes weeks to create a development environment. The way IT governance and policy has evolved over the last 10 year causes there to be too much cost, too much red tape, and too much politics. Another pain point is that the non-production environments are so different than the production environment. This causes the testing cycle to not be accurate and delays release (measured in weeks). Another important aspect of software development lifecycle is the speed by which software updates and changes are pushed out, an important aspect of this is versioning and maintaining the software in production is too difficult. Lastly, service availability and access to backend services (databases, security, ticketing, etc.) cause major issues to the agility and ability to pivot quickly. When you take all these aspects into consideration, the path to success is to not innovate or fail-fast.
Summary of the challenges facing devops shown in the figure below:
In the face of these challenges today, several obstacles come to light concerning Enterprise class devops. Many enterprises feel that devops is only for startups or 100% software focused products. This leaves many developers feeling that they are overworked already and their organization does not get it, so why push devops methodology. In addition, IT has put gates and process controls in place for good reasons, so why adopt a new process if the benefit is not obvious to the organization.
The main driver to adopt devops comes from the need to innovate and accelerate the business. Software is disrupting the industry and all enterprises are turning into software companies. The pace of change is not decreasing so it’s better to adapt sooner to this new methodology then latter. However to adapt devops, the short comings and challenges raised here need to be addresses. History is usually a good indication of how to address changes and challenges in the technology space.
As Cisco has embarked on the Intercloud journey to build the Platform for the Internet of Everything, we have had to tackle current development SDLC, devops, and future devops for the platform head on. We called our innovation Project Shipped and Shipped represents the beginning of the Platform for the Internet of Everything. What does this mean? Billions of devices need a platform to enable Rapid Development. However, flexibility is required to develop Private and Build and Deploy Everywhere. This framework is called Hybrid Devops. The result is no vendor/platform Lock-in and most importantly no compromises.
In addition, Shipped was designed to allow developers to:
- Develop Leveraging the Frameworks of their choice
- Build through CI/CD flow designed for multi-cloud and “bring your own” with consistent packaging and versioning
- Easy to deploy with service discovery and automatic service availability
- Manage you application and all services from a single interface across private and multi-cloud environments
- Team collaboration within organizations and across communities, built right into CI/CD
Shipped leverages the Microservices Infrastructure opensource project work I previously blogged about – Cisco Cloud microservices-infrastructure 0.2 released! – and the goal of Hybrid Devops is an easy to use experience for developers to Build, Deploy, and Run.
We have taken into consideration that there are other applications that your need to develop with so we have incorporated the ability to add Cisco Applications or connect to your private repository.
We understand that developers like to develop on their laptop so we initialize the development environment on your laptop, let you make a commit, and automatically kick off the build and test, and then deploy or publish.
Lastly, we understand the need to manage the application from this point on as agile application software that changes frequently with changing market needs and usage. The management of the application needs to be at the services level with all internal and external dependencies monitored around business objectives like Response Time, Throughput, and Errors. The ability to understand the application business objectives enables the business to manage the lifecycle and measure the value seamlessly.
We have a video of Project Shipped available here.
To get involved or to learn more, we welcome your contribution and have a landing page to help you get started: http://developer.cisco.com/shipped
Project Shipped is not just about Cisco, but a movement on hybrid devops and we are partnering with leading devops tools providers to deliver this unified platform including Hasicorp, Mesosphere, Cliqr, Nirmata, and Redhat. The initial release of Shipped has the following lifecycle and tools integrated and orchestrated end to end.
Please join us at Cisco Live! In San Diego (June 8th 2015), by participating in the pre-conference IoT and Cloud Hackathon (June 6th-7th), coming to the Devnet Zone for Theater sessions on hybrid devops, see demos of Shipped in the Cloud Devnet Zone, try out Shipped on Cisco Cloud using the learning lab, or attend the classroom instruction on Shipped.
IoT and Cloud Hackathon: Taking place on the Saturday and Sunday before Cisco Live. It’s free. Separate registration needed. Find out more and register for the Hackathon
The DevNet Zone page on the Cisco Live site brings all the information together, with a link to the different Cisco Live pass options. A DevNet Explorer Pass is only $49 which includes access to everything on offer in the DevNet Zone. Go to the DevNet Zone page on the Cisco Live site and register now!