Shifting to Software-Defined
Showing results for 
Search instead for 
Did you mean: 

The infrastructure does matter for DevOps


By guest blogger, Ka Wai Leung, Principal Product Manager – HPE CDI Solutions Product Management

In today’s digital economy, there is constant pressure to deliver innovations--especially for software-centric product offerings.  Software helps drive product differentiation and customer loyalty.  Example companies are Amazon, Uber, Netflix, Yelp, LinkedIn, Airbnb and Fidelity Investment.  For digital economy companies, software is their business.  Innovation means building, testing, and delivering quality software faster and more frequently by implementing strategies such as CI/CD (Continuous Integration and Continuous Deployment).

To be competitive, companies must reduce their time-to-market significantly. If these companies, especially those with mobile applications, have to wait six months or longer for IT to deliver the next big bang software updates, they will be at a significant competitive disadvantage.  To make their development and release process more agile, many IT groups are adopting the DevOps methodology so they can build quality applications at the speed demanded by their business.

When companies consider adopting DevOps, it can be a complex transformation process from their current development and deployment approaches. This can involve organization (people), process and technology changes-- essentially a culture change is required. This change can be summed up by doing the following:

  • Developing new approaches to integrate the development and the IT Ops sides closer together (and in some cases, allowing development teams to assume app deployment responsibilities)
  • Adopting new development, testing and release processes
  • Leveraging commercial and open source software tools that can facilitate CI/CD

In a DevOps world, the hardware platform is rarely discussed and thought to be irrelevant. Or is it???

Infrastructure such as compute, storage and networking are often considered as a pool of resources that can be provisioned and managed through APIs.  These resources can then be easily embedded into code and accessed throughout the DevOps lifecycle.

The hardware platform does matter for both the development side and the operations side of DevOps.  The right platform can help improve software release velocity and efficiency throughout the CI/CD life cycle.  Increasing the pace and quality of software deliver is the goal of DevOps.  As the code deploys to production, the right platform is essential to ensure a high degree of reliability, availability and scalability, along with end-to-end operations management.  Ideally, the Dev and Ops platform should be similar to minimize cross-platform issues from dev to production. Operations and lifecycle management are also easier.  Service providers, in particular, really understand this fact and employ homogeneous platforms whenever possible to minimize their OpEx and CapEx costs.

The recently released HPE hyperconverged platform, HPE Hyper Converged 380 (HC380) powered by Intel®, offers capabilities highly suitable for DevOps teams looking for a standard platform to host both their Dev and Ops environments.  The HC380 is a combination of HPE’s best-selling server (DL380) along with the HPE StoreVirtual software-defined storage and value-add software for lifecycle management, VM dispensing and monitoring.  It allows small to mid-sized businesses and enterprise remote offices to deploy VMs at cloud speed, simplify IT operations and reduce costs.


The HC380 offers these unique advantages for DevOps deployment versus generic servers:

  • Ability to create and deploy Dev, QA or Production VDI desktop in minutes to get DevOps team productive quickly and allow the team leads to control the types and versioning of the tools that get used, along with security auditing and source code access control.
  • Smart cloning technology that allows QA teams to quickly replicate large data sets or test databases in minutes without doing multiple full-blown installations. This will improve the functional testing velocity and lower overall CapEx since significantly less storage are needed for replicating data sets.
  • Five-nines of availability along with all sorts of built-in data protection, including replication, remote copy, and VMware certified multi-site clusters. While your Dev/QA environment does not have production data, no development/QA team wants to lose days’ worth of the latest source code, test scripts, and build artifacts and environments.
  • Two-tier storage (SSD & HDD) and adoptive storage optimization technology to automatically migrate frequent use data to the flash SSD tier, and less demanding data is moved to SAS or SATA.  This storage tiering will bring multi-fold I/O performance improvements where frequently accessed DevOps tools and the CI/CD pipeline can be served from the SSD layer.
  • Ability to start with a cost-effective two-node cluster and grow as your DevOps team grows. You can use the same architecture going from POC all the way to production.  And the HPE hyperconverged platform will give you linear scaling of compute and storage as incremental nodes are added.  Having a homogeneous environment between Dev and Ops ensures less system and virtualization portability issues as the application moves from Dev to Ops.

Using the HC380, an organization can greatly improve their development and QA velocity and total cost of ownership versus deploying generic servers. The HC380 will help DevOps team to be more productive more quickly, while improving the efficiencies of their CI/CD pipeline through the hyperconverged technologies mentioned above. 

In a DevOps world, the role of the Ops IT person is changing as the Dev and Ops teams are more integrated. The Ops person will need to be more application-centric and focus on the application performance, availability, integration and security.  For the Ops person to make this transition, he/she will need to focus less on the traditional infrastructure management tasks such as infrastructure provisioning, monitoring and life cycle management, and “up level” their skill sets.  This is where using a hyperconverged platform, such as the HC380 with its ease of management will help reduce the Ops related management time and enable the Ops person to spend more time on application-centric management.

For teams looking for a DevOps platform that they can grow into and have lasting value, the HC380 should always be a top candidate on their list. To learn more about the HC380 and how it can help your DevOps team, check out this solution overview.

Ka Wai Leung

Follow HPE Converged Systems

0 Kudos
About the Author