AWS Partner Network (APN) Blog
Unlocking Hybrid Architecture’s Potential with DevOps
Last week in our MSP Partner Spotlight series, we heard from Jeff Aden at 2nd Watch and learned about the value that next gen MSPs can bring to their customers through well managed migrations and through 2nd Watch’s Four Key Steps of Optimizing Your IT Resources.
Another area of new value that AWS MSPs can bring to their customers is management of their hybrid IT architecture, allowing customers at any stage of the cloud adoption journey to best leverage the AWS Cloud. This week we hear from Datapipe (APN Premier Partner, MSP Partner and holder of several AWS Competencies and AWS Service Delivery designations) as they discuss their approach and considerations in supporting their customers’ hybrid architectures.
Unlocking Hybrid Architecture’s Potential with DevOps
By David Lucky, Director of Product Management at Datapipe
Hybrid IT architecture, or what many customers call hybrid cloud, is increasingly prevalent in today’s fast-paced technology industry. Over the past few years, Datapipe has seen an initial reluctance towards cloud adoption transform into excitement, and hybrid architecture is emerging as a go-to solution for enterprise organizations looking for a way to manage their complex operations and run AWS as a seamless extension of their on-premises infrastructure.
Hybrid architecture gives organizations Application Programming Interface (API) accessibility, providing developers with programmatic access to control their environments through well-defined methods. APIs, commonly defined as “code that allows two software programs to communicate with each other,” are increasing in popularity in part due to the rise of cloud computing, and have steadily improved software quality over the last decade.
Now, instead of having to custom develop software for a specific purpose, software is often written referencing APIs with widely useful features, which reduces development time and cost, and alleviates risk of error.
With API accessibility, developers can easily repurpose proven APIs to build new applications instead of having to manage them manually. This gives them more room to experiment and innovate and creates a culture of curiosity. In this way, the API accessibility of hybrid architecture leads to a necessary rebalancing of development and operations teams looking to solve problems earlier and more automatically than was previously possible with purely on-premises solutions.
To maintain the culture of curiosity that’s enabled by API accessibility through hybrid environments, we recommend organizations remove the silos that traditionally separate development and operations teams, and encourage open communication and collaboration – better known as DevOps. Implementing a DevOps culture helps organizations take advantage of a hybrid infrastructure to increase efficiencies along the entire software development lifecycle (SDLC).
At Datapipe, we understand how critical the adoption of DevOps methodologies and agile workflows are for IT organizations to remain competitive and respond to the constantly evolving technology landscape. It’s the reason we expanded our professional services to include DevOps, and why we help organizations make the cultural switch to DevOps the right way, starting with people.
Individuals Over Tools
While many people conflate DevOps with an increase in automation tools, an organization can’t fully realize DevOps culture without starting with its people. A DevOps culture fosters open communication and constant collaboration between team members. It dissolves barriers between operations and development departments, giving everyone ownership over the SDLC as a whole, beyond their traditional, individual responsibilities.
Being able to see the big picture allows team members to transition from being reactive to being proactive. That, in turn, involves shifting away from addressing problems as they arise to determining the root cause of the problem and finding a solution as a part of a continuous improvement mindset.
Organizations that fully embrace this full-stack DevOps approach can provision a server in minutes instead of weeks, which is a vast improvement on the traditional SLDC model.
This mindset also means moving from a reactionary approach and solving problems through “closing tickets” to a proactive approach that involves consistently searching for inefficiencies and addressing them in real-time, so an organization’s software is continually improving at the most fundamental levels. Of course, addressing inefficiencies in the software also means addressing inefficiencies in workflows, which leads to the use of DevOps tools such as automation and writing reusable utilities.
However, productivity tools won’t increase efficiency on their own. An effective DevOps culture starts with open collaboration between team members, and then is reinforced by tools. At Datapipe, we see incorporating a DevOps culture through the lens of the “Agile Manifesto,” which promotes “individuals and interactions over processes and tools.”
When you combine agile working practices with DevOps, you can manage change in a feature-focused manner, providing faster interaction and response. Managing change in this way means that organizations achieve their goals through a strong DevOps culture that automates the majority of the overall development and delivery process, enabling teams to focus on areas that create a differential experience.
This takes time – and collaboration among team members – to set up. The real-time collaboration that marks a full-stack DevOps approach reduces the number of handoffs in a SDLC, thus accelerating the entire process, and decreasing an applications’ time-to-market.
Hybrid architecture growth is expected to continue. Industry analyst firm IDC predicts that 80 percent of all enterprise IT organizations will commit to hybrid architecture by the end of this year. This prediction is in line with what we’re seeing from our customers. As a next-gen MSP, we’ve seen an increase in enterprise companies looking for guidance on incorporating a DevOps culture to complement their digital transformations.
Take our work with British Medical Journal (BMJ), for example. BMJ started out over 170 years ago as a medical journal. Now, as a global online brand, BMJ has expanded to encompass 60 specialist medical and allied sciences journals with millions of readers. As a result of their dramatic growth, their old infrastructure could no longer support their application release process. In addition, as an increasingly global organization, BMJ’s capacity for allowing downtime – scheduled or otherwise – was diminishing.
To solve this problem, BMJ needed to move to a sustainable development cycle of continuous integration and automation, which is only possible through a shift to a DevOps type culture. We helped BMJ implement this culture while assisting with changes to their infrastructure.
The switch to a more open, collaborative culture not only allowed BMJ to implement a sustainable development cycle, complete with continuous integration and automation, but it also made them feel better prepared to take their next planned step of moving workloads to the AWS Cloud and embracing a hybrid environment.
If you’re interested in leveraging DevOps to get the most out of your hybrid environment, we recommend starting with the following considerations:
- Leverage object-oriented programming principles such as abstraction and encapsulation to build re-usable and parameterized components that can be assembled like building blocks. This can be done in configuration management with Chef Recipes, Puppet Modules, and Ansible Roles, or through infrastructure building blocks like Terraform Modules and AWS CloudFormation scripts.
- When automating infrastructure management, test destruction as deeply as the creation process. This will give you the ability to iterate and test cleanly.
- Balance the effort being put into upfront engineering versus operational management activities. More upfront engineering unlocks some great features with Auto Scaling on AWS. For more steady-state applications, the resources needed to set up and configure can sometimes be much less than the effort of working through automation. This makes it worthwhile to look for open-source modules to help you in your infrastructure and configuration management workflows.
- For Auto Scaling groups within AWS, consider, as you engineer your process, the time tolerances your workload has from the time when AWS detects the need for a new instance to when they are fully operational. Fully-baked Amazon Machine Images tend to be the fastest time to operational, but this would require building an image for every version of your application. Packer is a great tool for this purpose. In addition, the more you embed user data or configuration management processes, the longer your instance will take to reach an operational state. Finally, keep in mind processes like domain joins and renaming of instances, which require reboots, can add time to the launch process and use them as sparingly as possible.
- For a low-latency link between your resources in and out of the cloud, consider taking advantage of higher-level services like AWS Direct Connect, which provides a virtual interface directly to public AWS services and allows you to bypass Internet service providers in your network path. Datapipe client ScreenScape used Direct Connect to link their on-premises environment to Amazon CloudFront for a cloud environment that’s highly available, fully managed, and able to scale over time with proven capability.
Hybrid architecture offers organizations the power of both on-premises and cloud environments like AWS, giving them the tools to grow and innovate at a lower cost. For companies to fully capitalize on the benefits of these mixed environments, a culture change is necessary.
By shifting to a DevOps culture and enabling teams to work together in a full-stack perspective, organizations can not only increase efficiency in their SDLCs, but also open up opportunities for immense engagement and creativity – qualities necessary for innovation.
A next-generation MSP, with DevOps and Software-as-a-Service (SaaS) capabilities, can be a valuable guide for IT teams on their hybrid cloud journey. At Datapipe, we pride ourselves on being a next-generation MSP, and our proficiency with DevOps was a key differentiator that led to our position as a leader in the 2017 Gartner Magic Quadrant for Public Cloud Infrastructure Managed Service Providers, Worldwide.
By partnering with a next-gen MSP, like those included in AWS Managed Service Partner (MSP) Program, organizations don’t have to make the shift to DevOps on their own.