Ahana announces managed service for Presto on AWS

Ahead of the Linux Foundation's PrestoCon virtual event on September 24, Ahana, a relatively new challenger in the Presto query engine market, is announcing the release of its managed cloud service, Ahana Cloud for Presto. It's also announcing new seed funding.

The new seed funding totals $2.3M and is led by Lux Capital. This is in addition to previous seed funding of $2.25M, in June, 2020, led by GV, which also participated in the new seed investment. Ahana is also announcing that Vivek Bharathan, Ashish Tadose and George Wang of Uber, Walmart and Alibaba, respectively -- all contributors to the PrestoDB project -- are joining Ahana as Principal Software Engineers and co-founders. 

Also read: Ahana gets seed funding, and a tale of two Prestos

Funds and personnel

Customers in the Presto space will need to compare Ahana's funding and tech talent to that of Starburst, the current leader in the Presto market. Starburst has thus far received $64M over two funding rounds, and its tech leadership team includes the three people who created Presto at Facebook: Dain Sundstrom, Martin Traverso and David Phillips. When the three left Facebook, they established the PrestoSQL project, of which Starburst is now the chief commercial backer. Starburst is also a member of the Presto Foundation, which operates under the Linux Foundation umbrella and oversees the governance of the PrestoDB project backed by Ahana.

Also read: Starburst Data announces $42 million series B funding round

ZDNet was briefed on these announcements last week by two Ahana co-founders, CEO Steven Mih and CPO Dipti Bokar. The two explained that, as with many open source projects, complexity can be a challenge when it comes to deploying and running "vanilla" Presto. For example, generic Presto includes no database catalog, forcing use of the Apache Hive metastore or a compatible catalog platform like Amazon Web Services' (AWS') Glue, the two said.

Managing nodes in the Presto cluster requires knowledge and capabilities not unlike the rarefied skill set required to run Hadoop clusters. Mih and Bokar also said Presto's out-of-the-box performance can be lacking, and its indexing capabilities are modest, thus requiring manual optimizations, even for common queries. All of this creates obvious challenges for use of Presto on a self-service basis by analysts and data scientists.

At your service

Ahana Cloud for Presto has elimination of this complexity as a central design principal. It features a built-in metadata catalog; bundled storage (in the form of a dedicated S3 bucket for its AWS service); automatic deployment on Amazon's Elastic Kubernetes Service (EKS) using a supplied Kubernetes (K8s) storage operator, and requiring no specific K8s expertise; and built-in monitoring capabilities to maximize uptime.

Ahana Cloud for Presto is a Software as a Service (SaaS) offering in AWS Marketplace -- meaning it is not simply based on Amazon Machine Images (AMIs) that the customer must deploy itself -- and it runs by default in the customer's virtual private cloud (VPC). For this initial offering of the service, connectivity is focused on AWS data sources. Those sources consist of S3; the MySQL and Postgres offerings on RDS; Redshift; and Amazon ElasticSearch Service. Connectivity is not currently supported for Amazon Aurora, or the other RDS sources: MariaDB, Microsoft SQL Server and Oracle.

The connectivity offered at launch for Ahana Cloud is admittedly modest when compared with that of Presto competitor Starburst, which offers more than 30 supported enterprise connectors, several of them architected for parallelized query operation, according to that company. 

Two's company

But the important market dynamic here isn't whether these two competitors are at parity; instead, it's about the fact that there are two players of substance in the space. Going from singular to plural is a big deal, and arguably creates greater opportunity for both players, as it provides important additional validation of the market.

Ahana's offering also puts additional pressure on the cloud providers themselves: while Amazon's Athena service is based on Presto and while generic open source Presto is available as an option on Amazon EMR, AWS doesn't offer a standalone version of Presto, available on a pay-as-you-go billing model, capable of connecting to multiple data sources. Google Cloud only offers Presto as an optional component on its Cloud Dataproc service, and there's no first-party Presto offering on Microsoft Azure at all.

Presto change-o

While Apache Hive and Spark SQL once reigned supreme in the SQL-on-Hadoop (and thus SQL-on-data-lake) world, Presto seems within striking distance of taking the lead position. With Ahana Cloud for Presto announced for AWS, that momentum is now even greater.

Ahana Cloud for Presto is now accepting sign ups and early access will be available concurrent with PrestoCon on September 24. The company expects the service to reach general availability (GA) by the end of this year and says support for Microsoft Azure and Google Cloud will be added in the future.