Around the Storage Block
1748138 Members
3636 Online
108758 Solutions
New Article ๎ฅ‚
StorageExperts

How to avoid cloud lock-in for Docker containers with persistent storage

Want to avoid cloud lock-in with Docker, Kubernetes, OpenShift or Mesos? Whether you develop applications on-prem or with AWS cloud, the right persistent storage can help. Learn more now.

Cloud Containers_blog.jpgAs we head into the summer tradeshow season, itโ€™s time to reflect on what weโ€™ve learned from our customers. Over the past year, weโ€™ve met many more people who want to use Docker containers to make their applications portable to any infrastructure. Everyone wants to avoid cloud lock-in and thinks Docker containers can help.

Enterprises are quickly adopting Kubernetes, Docker and Mesos. AWS, Azure, and Google clouds have responded to this trend by delivering their own versions of Kubernetes-as-a-Service. So is packaging apps and their dependencies in containers enough to leave enterprises truly free to run those apps anywhere they want? The answer is no.

Why Docker containers wonโ€™t prevent cloud lock-in

Docker containers are extremely portable, but they arenโ€™t immune to cloud lock-in and switching-costs. Whether youโ€™re developing new cloud native applications or containerizing existing applications, here are the most common pitfalls that tie applications to specific clouds and some ways to plan for and address them up front.

The unexpected causes of cloud lock-in

Watch for these possible cloud lock-in scenarios:

Over use of cloud-specific Platform-as-a-Service (PaaS) capabilities or cloud APIs only available on one cloud: Such APIs enable rapid development and make life easy for developers, but they also make it difficult to move applications or components to other infrastructure later. Doing so requires costly application re-architecture.

Unexpected cloud data lock-in: Whether you choose to bring your own software-defined, file-system, or use cloud storage services directly, costly public cloud egress and bandwidth charges can make data expensive to move. And if your file system, database, or storage services arenโ€™t available on other clouds, you may not have a choice. As a result, data, containers and applications will tend to live wherever data is collected and processed. A few reasons why this is costly:

  • Itโ€™s easy to consume 4-to-5x more capacity than you would to store the same data on-premises because public cloud storage lacks basic data reduction and efficiency features. [i]
  • Unlike compute, which can be turned on or off, data costs are steady and only increase with time โ€“ even if you donโ€™t access the data.
  • Inefficiencies of public cloud storage can increase application development time and costs.

Best practices for addressing cloud lock-in with design considerations up front

Keep these two things in mind:

1. Decide when you will and wonโ€™t use cloud-specific APIs and stick to this policy:

  • Consider only using them to develop applications and components youโ€™re sure you want to commit to one given cloud.
  • Do some โ€œwhat ifโ€ analysis upfront and be ready to pay changing costs as you grow and scale.

2. When designing new applications, consider where you will store digital data and files up front and estimate all costs up front:

  • Estimate and monitor ongoing capacity and data management costs. Since cloud costs change and applications grow, do some โ€œwhat ifโ€ calculations to plan for those scenarios up front.
  • Always have an exit. Consider: How much will it cost to move and share data between clouds and between cloud and on-prem using your system?
  • Donโ€™t forget bundling and packaged costs for related services. If you need to move up or down on CPUs can you do that without upgrading the storage and vice versa?
  • If youโ€™re considering a BYOD strategy, factor in the cost of additional staff or professional services to manage another storage island for your applications.
  • Plan the most carefully for cloud-specific storage or database-as-a-service offerings because they lead to both cloud API and data lock-in for applications.
  • Look for the self-service automation development teams need.

An even easier way to address these concerns

Consider designing applications to use container-integrated, enterprise-grade cloud storage that will help you:

Cut costs

  • Cut capacity consumption by up to 80% with enterprise storage efficiency.[1]
  • No predatory bundling costs since you can scale compute and storage independently.
  • No need to hire more specialistsโ€”when you can rely on enterprise class support.

Avoid lock-in

  • Use AWS and Azure Cloudโ€”without the egress charges.[2]
  • No costly data migrations with affordable multi-cloud application and data portability.

Speed up DevOps and CI/CD

  • No need to take time away from production with advanced data services for DevOps teams.[3]

HPE was the first to deliver advanced data services for enterprise DevOps teams working with existing applications, the first to provide rich Kubernetes and OpenShift integration for both of our lines of all-flash arrays, and the first to announce a true enterprise-grade multi-cloud storage service for AWS and Azure with HPE Cloud Volumes.  

See you soonโ€”at DockerCon 18 and HPE Discoverโ€”where you can see innovation in action

We continue to be first with valuable innovations for customers. Visit us at DockerCon Booth #S9 and HPE Discover demo #419 to see our newest innovations, enhancements, and tech previews.

Container focus:

HPE Nimble Storage Tech previews:

  • Docker EE on AWS leveraging advanced data services with our HPE Cloud Volumes Docker Volume Driver for data intensive CI/CD pipelines
  • Replicate on-premises Microsoft SQL data to a persistent volume in AWS cloud for dev/test

MeStacey Gilmore HPE Storage.jpeget Around the Storage Block blogger Stacey Gilmore, Containers Solutions Marketing. A former HPE DevOps Engineer, Stacey is surprised and delighted to be back where she first began her career: at HPE.

 

 

[1] Most enterprises storage platforms deliver anywhere from 5x data reduction or more compared to cloud storage. A few examples: HPE Nimble Storage delivers up to 5x data reduction and Reduce your storage capacity requirements by 75 percent: HPE 3PAR Get Thinner Guarantee

[2] Avoid cloud egress charges with HPE Cloud Volumes

[3] BrightTalk: Stateful containers for DevOps and Beyond

 

 

 

0 Kudos
About the Author

StorageExperts

Our team of Hewlett Packard Enterprise storage experts helps you to dive deep into relevant infrastructure topics.