Fivetran targets data security by adding Hybrid Deployment

Hybrid Deployment

Enterprise data is complex. It comes from myriad sources in an array of forms and contains different types of information.

Some of it is simple and straightforward. It can be ingested from its original source and moved into a system on a public cloud where it can then be prepared for analysis. It needs to be kept secure because it often contains proprietary information that if leaked could lead to a competitive disadvantage for an enterprise. But it doesn't contain data that, if accidentally exposed, could violate privacy regulations or result in other harm to an organization.

Other data is more sensitive. For example, it could contain personally identifiable information that if exposed would lead to a regulatory violation.

As a result, enterprises often keep such data in their own environment so it never gets moved -- even through secure channels -- into an external system such as a public cloud or a vendor's platform. Private clouds are one method of keeping data more secure.. On premises databases are another.

To meet the needs of those organizations subject to greater regulatory oversight than most others or that deal with large amounts of sensitive information, established data integration vendors such as Boomi, Informatica and SnapLogic provide hybrid deployment capabilities.

Meanwhile, Fivetran is not long out of its startup phase, Henschen pointed out.

"Fivetran is a fairly young, cloud-focused company," he said.

As a result, though it competes with mature vendors, its most similar peers are other data integration vendors still attempting to mature, such as Airbyte, DBT Labs and Matillion. Now, as it attempts to grow, Fivetran is adding certain capabilities that more evolved vendors already provide.

Hybrid Deployment, therefore, represents Fivetran's own effort to meet the needs of organizations subject to strict regulatory oversight or that deal with large amounts of sensitive data. Such organizations often struggle to develop systems that enable data-informed decision-making, the vendor noted.

With Hybrid Deployment, Fivetran is supplying them with the means to integrate data without it ever leaving their secure environment in what Kevin Petrie, an analyst at BARC U.S., called a logical next step in its development.

"Fivetran first distinguished itself as a simple, cost-effective way to move data from many SaaS applications in the cloud to cloud-based data warehouses such as Snowflake," he said.

Later, in 2021 in conjunction with raising $565 million in funding, Fivetran acquired HVR, adding capabilities that enabled customers to extract data from on-premises databases, Petrie noted.

"Their new hybrid capabilities extend this strategy by enabling Fivetran customers to manage and govern data pipelines on premises as well as in one or more clouds," he said.

While Hybrid Deployment is an evolution for Fivetran, providing enterprise-grade capabilities that its more mature competitors also provide, the impetus for its development came from customer feedback, according to chief operating officer Taylor Brown, a Fivetran co-founder.

Certain customers use Fivetran for some of their data integration and movement. Without being able to manage their data in their own environments, however, they were unable to use Fivetran for all their data movement and integration needs.

The security fears that cause certain organizations to avoid using public clouds to manage data are not unfounded, Brown noted. For Fivetran to serve their needs, they required something more from the vendor.

"These customers might use Fivetran for most data sources but have to rely on cumbersome DIY or self-hosted solutions for the most sensitive of sources," Brown said. "Fivetran's Hybrid Deployment … provides them with the flexibility to maintain full control over sensitive data while benefiting from the ease and scalability of a managed service."

Hybrid Deployment provides Fivetran users with a single control pane from which they can manage all data sources, whether they are in public clouds, private clouds or on premises.

Intended benefits include the following, according to Fivetran:

• Full visibility by enabling customers to monitor all their data pipelines from a single interface in their own environment rather than Fivetran's.
• Security features that let users set access controls, mask sensitive data and track data as it moves through pipelines to ensure that it remains compliant.
• Compatibility across multiple environments including on premises, private clouds and public clouds such as AWS, Google Cloud Platform and Microsoft Azure.
• Cost control by providing customers with detailed reporting tools and enabling them to track their usage.
• Pipeline scalability and customization to meet an organization's unique needs.
• Simple setup.

While all are valuable, compatibility across environments and cost management are potentially the most significant for Fivetran customers, according to Henschen.

Independent vendors such as Fivetran tend to address compatibility better than tech giants that provide data integration tools as part of larger data management offerings, he noted. Meanwhile, cost management measures are critical as more workloads are moved to the cloud and cloud spending increases for enterprises.

"All [the intended benefits] are important and ones that many integration vendors try to address," Henschen said. "The ones that stand out to me are compatibility across environments … and cost management, which is really crucial to all customers but not all vendors address the need adequately."

Petrie likewise highlighted the cost control features as particularly notable. Being able to accurately predict cloud spending is critical for enterprises, especially as they increase their investments in generative AI, traditional AI and machine learning fueled by technological advances such as the evolution of large language models and added compute power.

"I like their ability to monitor utilization and consumption of pipeline resources from a cost standpoint," he said. "This supports FinOps initiatives in which enterprises aim to predict, measure, optimize and govern cloud-related costs."

Looking ahead

With Hybrid Deployment now part of Fivetran's platform, the vendor's aim is to continue adding features that ultimately make access to data as simple and reliable as possible, according to Brown.

Specifically, the vendor's roadmap includes adding tools that provide more support for customers as they develop AI and machine learning workloads, including further addressing data security and compliance.

One such tool will combine the capabilities of Hybrid Deployment and Fivetran Managed Data Lake to enable any organization -- whether dealing with large volumes of sensitive data or not -- to use their unstructured data to help train AI models and applications.

Unstructured data, often stored in data lakes, now makes up more than three-quarters of all data and is key to training AI models and applications. Unstructured data, however, is not easy to integrate with structured data to provide enterprises with a complete view of their operations.

"As AI workloads and applications continue to be top of mind for customers, we'll be building more functionality into the platform to support those efforts," Brown said.

Eric Avidon is a senior news writer for TechTarget Editorial and a journalist with more than 25 years of experience. He covers analytics and data management.