Skip to main content
Exploring what’s next in tech – Insights, information, and ideas for today’s IT and business leaders

How to optimize databases for a hybrid cloud world

Modern, multi-tier applications can be on public and private clouds and on premises, with databases on any or all of them.

Enterprises run hundreds or thousands of applications, leading to database sprawl that only gets worse over time. Public and private cloud solutions let organizations consolidate infrastructures, applications, and databases onto shared hardware and software-defined infrastructure to effectively manage the hardware and software platforms, avoiding silos.

A modern hybrid cloud database includes workloads that can be moved, managed, metered, and optimized in either public or private clouds, or even on premises. As enterprises move their default databases and management to a hybrid cloud combination, a number of considerations for developing a data or database architecture arise.

The potential benefits of a hybrid cloud database environment include:

  • Flexibility and agility
  • Elasticity and cost optimization
  • Self-service capabilities
  • Faster delivery of data
  • Avoidance of deployment bottlenecks
  • Ready access to the latest technology

As new applications move to the hybrid cloud, data and analytics capabilities will need to follow. Business initiatives such as digital business transformation will require a greater diversity of data and analytics capabilities; they'll need the flexibility and agility a hybrid cloud database platform provides. In a multi-tier application stack, the presentation service could be on a public cloud, the application service might reside on a managed private cloud, and the database service might reside on premises. Increasingly, the database layer can exist on any or all of these.

Please read: Start making sense: Building modern data platforms

In a hybrid world, every application aims to be as nimble as possible, often adopting an as-a-service approach, reducing capital investments in infrastructure and data centers. Designing data centers to accommodate peak loads, with infrastructure sitting idle much of the time, is far costlier than adding cloud resources when needed to accommodate peak periods.

A well-designed hybrid cloud database can allow both developers and line-of-business managers to gain access to IT infrastructure and services through a self-service portal. Access to services in this manner reduces the burden on IT and greatly increases the agility of the business operation by promoting the delivery of new products and services more quickly. It eliminates the barriers that slow down business and development teams. New sources and types of data become easier to create and deploy, while developers and test engineers can better access the resources needed.

Four key aspects of database modernization

Major considerations for database modernization include:

  • Capacity planning: The process of right-sizing database instances at scale on either a private or public cloud platform is always based on performance metrics measured in time series.
  • Transformation needed: This applies to both the application and database levels. Application and SQL modernization and migration to the cloud or hybrid cloud is the process of redeploying an application with the goals of financial and operational efficiencies and increased scalability and resiliency.

Each transformation approach will impact the application stack at different levels, since each approach presents new challenges and cost implications to obtain all the benefits of cloud services:

  • Grouping applications and databases into logical units. Developing practical and effective waves of database migration typically includes a methodical and iterative approach. Scoring applications for migration involves objectively applying a set of rules against application components, clusters of applications, and so on. The preferences of the business or client are, of course, paramount.
  • Load balancing, high availability, and disaster recovery. High availability, by definition, permits a database to continue to function even when major components fail. Mission-critical systems cannot tolerate interruption in service, and any downtime may cause damage or result in financial loss. Ensuring that databases can fail gracefully and restore operations without disruption of service is crucial.

Which database, which hybrid platform

The process of right-sizing database instances at scale on either a private or public cloud platform is always based on performance metrics measured in time series.

The process for right-sizing database instances at scale based on on-premises performance metrics involves using scripts (OS- and SQL-based) to collect the metrics from on-premises databases and use for data analysis, leading to recommendations regarding database instance size. This solution works for the instance sizing, with the scope ranging from one to many databases.

Lowering the cost of database management is a primary driver in moving to a public cloud. Eliminating capital expenses for hardware and software, along with the operating expenses of installing, maintaining, updating, and patching databases and additional administration overhead, are balanced against the operating expenses levied by a cloud service provider (CSP). Your return on investment must also include the conversion and migration costs.

Please read: Successful hybrid cloud projects require a detailed roadmap

Reliability and redundancy are important considerations for cloud adoption. Running a commercial cloud platform means tens or hundreds of data centers worldwide, with high-reliability service-level agreements. Care must be taken to properly deploy high availability for databases, since even CSPs have regional or data center outages that need to be covered by replicated copies, redundant zones, and similar strategies.

The ability to dynamically scale up or down is a compelling reason to deploy a database in the cloud. Traditional on-premises database deployments may implement stretch database approaches where designated segments of the database remain on premises, while others are stored in mirrored environments in a CSP to balance loads, offer excess capacity during peak periods, or replicate for high availability and disaster recovery purposes. This is a classic example of a hybrid cloud database.

Security under a CSP is typically stronger than that on premises. CSPs have armies of people tracking security bulletins, conducting penetration testing, and related activities.

Still, whether based on perceived security exposure, regulatory and compliance considerations, or sunk cost in existing infrastructure, there remains reluctance to move to a CSP. A private cloud deployment, where the above features are offered on premises, is a popular alternative. If a CSP's on-premises layer is deployed, provisioning and migrating can be intermixed in a hybrid model.

What to look for in a database platform automation solution

An automated deployment process minimizes the chance of errors and failures. The release process for database change should enable updates with minimum disruption of service. Versioned release into the production environment allows for simultaneously switching between the new version and previous releases, should a previous release fail.

A database platform automation solution is elastically scalable, providing on-demand access to database services using a self-service portal. A self-service portal integrated into the common control plane enables the on-demand access of infrastructures and the different platforms to meet business needs. Such a solution provides accelerated time to value with simplified, self-service provisioning in minutes.

Please read: A data fabric enables a comprehensive data strategy

Standardized processes provide rapid, consistent deployment and lifecycle management, maintaining security, compliance, and data sovereignty, while controlling sprawl, reuse of infrastructure, and software licensing cost.

Why you need automation

Key strengths of a successful database platform automation and management solution include:

  • Database provisioning of single or multitenant instances on bare metal and hypervisors (like VMware) with all the best practices applied.
  • Ability to scale compute and storage resources up or down on demand, including add node, delete node.
  • End-to-end database lifecycle management, including patching, backup and restore, cloning, and upgrades.
  • Integration with DBMS vendor patching and provisioning to keep the environment evergreen using golden images.
  • A workload profiler for better consolidation and DBMS license usage metrics to optimize future support costs.
  • API-based modules that provide seamless integration with your existing infrastructure and control plane solution (such as Terraform and ServiceNow).
  • Support for a wide range of server, storage, and network infrastructure configurations.
  • Integration with backup and restore for data protection.

A flexible, scalable solution

Addressing database sprawl and lowering the cost of database operations is top of mind for most organizations, particularly since getting control of it allows for a focus on innovation. Simplifying database service provisioning allows database operators to focus on high-value tasks. Reducing acquisition and operating costs of database platforms improves ease of use and responsiveness, and simplifies management.

A successful database platform automation solution brings together compute and storage technologies along with software stack capabilities and consulting and support expertise. It's offered in an effective as-a-service model. The time is now for a flexible, scalable, cloud-like approach to database platform automation.

This story has been excerpted from the author's e-book Modern Database Architecture in a Hybrid Cloud World.

Lessons for leaders

  • Multi-tier applications in a hybrid cloud environment enable high performance, even with vast workloads.
  • Modern hybrid cloud systems let you auto-tune databases and applications based on performance metrics.
  • The complexity of databases in multi-tier applications requires platform automation if quality and security problems are to be avoided.

This article/content was written by the individual writer identified and does not necessarily reflect the view of Hewlett Packard Enterprise Company.