How to Choose a Cloud Data Management Architecture

It is crucial that D&A managers make the right choice of cloud data architecture for their organization.

Donald Feinberg, Vice President and Analyst Emeritus of Gartner’s ITL Data and Analytics (D&A) Group, explores the different types of cloud architectures for data management and explains why D&A leaders must balance the risks and benefits of each

The need and use of data, whether customer or business data, is becoming more and more beneficial to organizations today. It helps businesses stay competitive and stay ahead with intelligence to make smarter decisions, faster.

However, it’s important to recognize that a data-driven strategy can demand too much of a business, especially if the right tools and solutions to handle these additional needs aren’t in place.

Solutions such as cloud data management architecture are therefore essential. However, D&A managers should be aware of the different architecture choices – from on-premises to multi-cloud and cross-cloud. They should understand the risks and rewards of managing data in diverse and distributed deployment environments.

Here, I look at the different cloud data management architectures and considerations D&A managers need to take into account.

1. On-premises to cloud

In a local-to-cloud (also known as ground-to-cloud) model, various components of an application architecture can reside on-premises and/or in a cloud. Database management systems (DBMSs) can reside on-premises and applications that connect to them can reside in the cloud, for example, a Business Intelligence (BI) dashboard application.

There are two variations of on-premises to cloud architectures:

  • Active (formerly “hybrid cloud spanning architecture”)
  • On-demand (formerly “use-case specific hybrid cloud”)

An active approach, as the name suggests, deals with the active management of data between the two environments. This can include architectures with data residing both in the cloud and on-premises, such as the ability of the DBMS to have replicas, partitions or shards residing on-premises and others in the cloud for the same database. data.

There are many application use cases for this type of functionality, including: partitioning data by age, access frequency, or geography; dynamic capacity allocation to respond to inconsistent and sudden demand for resources; and regulatory requirements governing data locality.

In an active on-premises-to-cloud model, it is essential to understand the characteristics of the data flow (for example, whether data is entering or leaving the cloud and the expected data volumes). There may be issues with latency, i.e. the time it takes to move data between on-premises and the cloud. Additionally, there may be financial implications related to CSP data egress charges. Integration, metadata, and governance practices spanning multiple environments should also be considered. Service level agreements (SLAs) must be defined and tested. This can lead to the requirement for a special communication link between on-premises and cloud components, resulting in greater cost implications.

In an on-demand approach, the components remain separate. Data is moved between environments only when necessary to support business activities such as disaster recovery planning or development lifecycle functions. For example, all development, test, quality assurance (QA), disaster recovery (DR), or production instances of a DBMS can reside on-premises or in the cloud. Although financial and latency considerations are still important, compatibility is the primary concern in this scenario. Many organizations may not be comfortable with less than 100% code compatibility between cloud and on-premises environments, which limits cloud service provider (CSP) selection to those who can meet these rigorous requirements. .

Key considerations for on-premises to cloud deployments include: moving data both in volume and direction (active); and component compatibility between environments (upon request).

Related: “On-prem first” – a hybrid solution for workflow storage – enabling organizations to take a hybrid approach, while centering on-premises application storage for security and reliability

2. Multicloud

Multi-cloud models integrate one or more services from multiple cloud providers (and may optionally include on-premises or hybrid architectures). In this scenario, the difference is that services from multiple cloud providers are used. A DBMS offering and the applications that depend on it can be deployed both on-premises and/or on one or more clouds.

As such, all hybrid cloud considerations may apply along with the additional considerations of deploying software across multiple cloud environments. These offerings have historically been limited to independent software vendors (ISVs) rather than native CSPs, because ISVs have a greater incentive to ensure their software works in as many environments as possible. However, cloud service providers are increasingly engaging in multi-cloud and cross-cloud scenarios.

The multi-cloud scenario usually attracts end users who are worried about dependency on one cloud provider and want to be able to easily move their applications to another cloud provider. By providing a semantically compatible offering that runs identically across multiple clouds and on-premises, multi-cloud-enabled DBMSs promise easier (though still not easy) migrations, as the primary concern will be data migration, and not the rewriting of applications.

On the contrary, for multi-cloud deployments, it is paramount that D&A managers consider the compatibility of components between environments and the different cloud capabilities for provisioning, management and governance.

Related: Why Enterprises Should Embrace Multi-Cloud — Neil Templeton, VP, Digital Innovation Marketing at Console Connect, explains why enterprises should embrace Multi-Cloud

3. Intercloud

Intercloud refers to the active management of data across multiple clouds. In a cross-cloud model, different components of an application architecture can reside on different cloud platforms and exchange data. For example, Microsoft’s PowerBI can connect to a Salesforce database that resides outside of Azure cloud infrastructure.

Intercloud models are less commonly used today. At the same time, they are increasingly of interest to those looking for more advantageous pricing models, specific tools not available from other CSPs, risk mitigation through the use of multiple CSPs, and meeting requirements data sovereignty through diverse data localization. For example, regulatory requirements may prohibit data from residing outside a country’s geographic borders.

For cross-cloud deployments, D&A managers need to pay particular attention to the movement of data, both in volume and direction.

Written by Donald Feinberg, vice president and analyst emeritus of Gartnerof the ITL Data and Analytics (D&A) group

Comments are closed.