Cloudera execs from its CEO to its SVP of Product Management, share their perspectives with frank comments about cloud adoption, customer use cases, partnerships, coopetition, and how prioritized investments are helping enterprises achieve their cloud goals.
Cloudera’s Chief Revenue Officer, Frank O’Dowd described cloud consumption as an important measure of the ability to take advantage of the technology and move forward with it. “Cloud consumption means you are using the cloud platform that you made a purchase. And our growth rate right now in cloud consumption (at 42%) is higher than that of all our competitors.”
This is no mean feat considering that 100% of Cloudera’s customers were onboarded this year.
Later, CEO, Charles Sansbury, shared with the EVOLVE24 keynote attendees, where they are going with this, when he said that hybrid is the way to go. This isn’t just a mantra, but also a realization among CIOs as they complement their public cloud workloads with on-premise ones.
“In the past 12 months, I think it has become generally accepted that companies will not just maintain their on-premise capabilities, but most generative AI at scale, will be run in on-premise hardware for purposes of security, scalability, and ultimately cost.”
Ultimately, the organization wants to help its customers achieve their cloud goals, and it is still delivering private cloud solutions that they are investing more into to take advantage of the momentum. Frank had said, “Our private cloud data services are growing 119% year over year, and we expect this number to grow significantly more this year.”
Move the model, not the data
What this has meant also is not having to bring a customer company’s data to the cloud to run on AI models. Instead, these models can be brought to the data. Charles explained, “So, let the data reside as it is, put it in a more modern data architecture like the data lakehouse, use knowledge graphs, and use table formats to create amalgamated sets of data.”
Also, these data sets would be able to pull from data silos thanks to the open data lakehouse having federation capability across multiple data siloes to create a trusted set of data.
The organization’s unique advantage of having worked with many world-leading companies, sees them being able to help customers with a “recipe book of industry-specific use cases driven by everything from machine learning models to generative AI.”
With a reference framework of how companies have solved problems, they can speed up identification of an end solution for companies that face similar problems.
“And this is such a big focus for us, not just to help you with your technology foundation but also to share with you, some of what we have learnt from a business perspective,” Charles explained.
Operationalizing AI for enterprises
Charles shared how Cloudera has aligned itself towards helping customers with their hybrid clouds as well as with operationalizing their enterprise AI. “I hope you will see we have prioritized investment across hybrid deployment capabilities across the components of modern data architecture, with primary focus on the open data lakehouse, REST catalog, and table formats like Iceberg.”
He named Verta, a recent Cloudera acquisition with a history in LLM management for companies. “With this, we are allowing you to empower your data scientists to manage performance of models, give you a framework for evaluating performance of different models, and help you decide which to put into production.”
True hybrid is the way to go
CIOs have accepted hybrid cloud as the way forward, but to really glean value, true hybrid is the way to go. Venkat Rajaji, Senior VP of Product Management, said, “I’ve been hearing our customers are moving to Iceberg as a place where they are going to standardize all of their data across their (IT) estate.
“We’ve been supporting it since 2020 on all of our engines. Every single one supports read and write to Iceberg today, table format, and the REST catalog… we can now serve up that interoperable ecosystem for our customers.”
According to Venkat, as long as an engine, not necessarily Cloudera’s; that supports REST APIs is used, the user can access data from Cloudera’s open data lakehouse. This means enterprises can leverage their existing solutions and maximize their investments in any analytical tool they would like to use.
Venkat frankly said, “If the hypothesis is true, and more customers centralize and standardize on Iceberg, then our lakehouse optimizer becomes absolutely critical.”
All of these components line up with Cloudera’s data mesh vision which Venkat admitted he is excited about.
Cloudera’s Chief Strategy Officer, Abhas Ricky had shared, “You want to be able to get to a point where the data never leaves your enterprise, hence the Enterprise AI Ecosystem to service this.”
This requires a series of capabilities and Cloudera is working with best-in-class providers to enable customers with this capability. With Cloudera’s mantra literally being “True hybrid is the way to go,” it wants to empower much required interoperability for everyone without them needing to pay an integration tax.
Abhas explained why their partnership with Snowflake is what the industry has been looking forward to.. “One very real example is a very large customer and their major airline. They were using us for our metadata governance, and Snowflake for fast query.
“Now, with this partnership, they are super excited, and they are like, ‘Hey, we want to work with you to go forward with this, because the interoperability will be super helpful.”
Another example is a large bulge bracket bank, or one of the world’s largest global investment banks, that Cloudera is excited to be a design partner for, and “they are gearing to get through with that,” Abhas said.
Data center APAC activity
Recently, APAC and especially Malaysia has been a hive of activity with hyperscalers and even one large database player, Oracle, investing to build data centers and deliver better outcomes for their customers in the region. How will all this pan out for the AI and data analytics ecosystem?
At the end of the day, we are not an infrastructure company. We are a software company, so we don’t care where the workload sits on.
Abhas observed, “A majority of the hyperscalers, cloud providers have data centers, but they also have software services on top of that.”
Recognizing the coopetition element that exists between the hyperscalers like Microsoft, Google, AWS, and Cloudera, Abhas said, “The idea is we will continue to work with them. Our software products around AI, data science, ML ops, LLM ops, and so on, might compete against some of theirs, but at the end of the day, we are not an infrastructure company. We are a software company, so we don’t care where the workload sits on.
Brandishing a multi-cloud strategy, the organization believes one of its core strengths is no vendor lock-in for large Cloudera customer companies as they typically use Cloudera software on two, or three, or multiple clouds.
“One of the core things people want is scaled AI workloads on top of GPUs in environments whereby they want to get to a better TCO, whether GPU infrastructure layer is being provided by NVIDIA, or somebody else that is the customer choice.
“We provide that flexibility for deployment form factors for GPUs. At the end of the day the name of the game is for customers to use and build AI applications that are faster, cheaper, and that’s where we’re at,” concluded Abhas who also observed APAC as one of the fastest growing regions with the most innovative use cases that are driven by Cloudera customers themselves.