Connected clouds on a circuit board.
Image: kras99/Adobe Stock

Snowflake does data. The company “does” data-centric software development at the application level with its Native Application Framework; it facilitates portal-level data services trade and exchange with its Snowflake Marketplace offering; it offers platform-level data management with its wider approach and self-styled stance as the data cloud company. That’s doing data at several levels.

Drawing the year to a close with some relatively substantial product updates, Snowflake used its Snowday 2022 event to stage the San Francisco leg of its Data Cloud World Tour and explain where its toolset is going next.

SEE: Hiring Kit: Cloud Engineer (TechRepublic Premium)

Described as now benefiting from “transparent” performance enhancements across its single compute/data engine — transparency here meaning power and functionality advancements that Snowflake itself has engineered organically within its own stack as well as taking advantage of improvements in new hardware types — Snowflake insists its core data cloud offering is now faster and more performant.

The promise to business is a route to enable organizations to mobilize their data to run their most important workloads.

Cross-cloud capability

Looking at Snowflake’s latest enhancements, the company has clearly recognized the cloud diversity landscape for what it is. Organizations of any reasonable size using cloud computing technologies (in increasingly cloud-native deployment models) are using more than one cloud instance. Spreading risk, cost, functionality and governance compliance across more than one cloud services provider to embrace multi-cloud is today regarded as plain and simple common sense.

Spreading the cloud estate further to keep a portion of compute and storage on-premises in a private cloud data center is also a key part of this sensibleness, which we now call hybrid cloud. Fanning out an organization’s cloud instance stack to spread some individual application workloads across clouds and create what we can now call poly-cloud is also equally sensible, but perhaps two or three times as complex.

To reflect this reality, Snowflake is advancing its cross-cloud capability. Known as Snowgrid — what icy analogy can’t the firm find a snappy product name to align to? — this is a technology layer that interconnects clouds and regions and powers Snowflake’s collaboration, data governance and business continuity capabilities.

As organizations continue to scale their data needs, the ability to work with various teams, regions and clouds across the globe is imperative. Snowgrid is built to empower organizations with one connected experience across globally distributed teams and businesses, all of which work to unify governance in order to comply with regulations. In basic terms, it enables organizations to replicate more than just data for increased business continuity.

“Our success is predicated on our customers being successful when using Snowflake, which is why we continually strive to deliver improved performance and economic value for customers,” said Christian Kleinerman, senior vice president of product at Snowflake. “With one single product, we can innovate faster and deliver improvements that enable our customers with a unified experience across clouds and regions, so they can protect their data at cloud scale, operate more efficiently and collaborate globally.”

What Snowflake does, actually

To explain what Snowflake actually is, what Snowflake actually does and how Snowflake is actually deployed, we can look at how the Snowflake Application Framework is used by software developers to handle the data workload responsibilities associated with an application.

To create what is typically a data-intensive application working with terabytes and petabytes of information traversing complex multi-tiered data pipelines, Snowflake offers services such as “stored procedures” (code to control database functions) to provide packaged and tested routes to building efficient data-centric application architectures.

The company is known for its single elastic, cloud-flexible performance engine that accelerates common queries, the cadence of replication speeds, the creation of new use cases and the ability to target latency issues. Among the most prized possessions in its own technology stack, the elastic engine gets a lot of R&D to drive what it calls continuous innovation.

Snowflake’s latest release continues the company’s trend for driving performance enhancements. The company has also worked to provide new visibility into the impact of performance optimizations.

Improvements include a Query Acceleration Service, now in public preview to speed up outsized data queries by providing a burst of additional resources without needing to scale up the overall compute. Think of it like a Netflix user who only ever downloads half-hour episodes of “Curb Your Enthusiasm” every night who then one day suddenly opts to stream The Godfather trilogy in full. If it’s an infrequent burst, it wouldn’t be fair to try and make someone adopt a new contract. Users appear to appreciate the balance between power and flexibility.

“DoorDash has continued its growth in terms of product innovation and geographical distribution over the last few years,” said Vaibhav Jajoo, head of data engineering & BI at DoorDash.

With its growth, Jajoo says that his firm’s data volumes have exploded.

“We are collecting new data points, ideating and experimenting on new business ideas and generating new insights,” Jajoo said. “We were looking for a scalable data infrastructure partner who can scale and grow with our growth. Snowflake has turned out to be that partner for us. Snowflake has been able to work with us to scale to our data needs, and address our complex business requirements.”

Snowflake Marketplace data listings

A key part of why Snowflake is doing well in the data marketplace war is its ability to offer so-called “data listings,” and within the Snowflake Marketplace, there are two types. A free data listing provides access to a published data set from another Snowflake user. This would typically be generic, aggregated or non-customer-specific data.

A personalized data listing allows Snowflake Marketplace users to request specific data sets from other Snowflake users. This can be premium data that a provider charges for or data that is specific to each consumer.

In both cases, each listing includes details about the shared data, sample usage examples and information about the data provider.

Snowflake is now advancing cross-cloud collaboration to provide customers with increased flexibility when sharing data listings. Providers can share listings publicly or privately with global business ecosystems. Providers can also choose to make their data automatically available and actionable for consumers across supported regions and clouds.

Also here in this winter update, we find cross-cloud data governance positioned as a means of simplifying governance at scale, which is a service that clearly helps organizations comply with international regulations and enjoy a level of unified visibility across their data ecosystem.

Cross-cloud continuity

Finally, let’s also point to the company’s efforts intended to simplify global operations for better cross-cloud continuity. Snowflake enables customers to replicate data across clouds and regions, unlocking greater levels of resiliency and ensuring minimal business disruptions in the event of a disaster.

If the non-data scientist and software developer layperson were to catch a whiff of this discussion but not really follow what’s going on, would a layperson’s elevator pitch help to distill the emotion behind some of the technology updates being tabled here?

This is essentially a story about how information is used in, through, across, beneath and above the cloud-centric applications that we all use on our devices every day.

To make functional connected data move between raw development environments, formalized secured databases, chaotic unstructured data lakes, ordered data warehouses and frenetically-charged data marketplaces, we need a means of currency, a market square, a set of by-laws and perhaps a town crier. Snowflake aims to be all of that and perhaps more.

By Umiddoc