Ahana Goes GA with Presto on AWS
Organizations that want to analyze massive amounts of data stored in S3 and other cloud repositories may be interested in Ahana, which today announced the general availability of the first cloud-native managed service for Presto on AWS.
Presto is a SQL query engine originally developed at Facebook as the follow on to Apache Hive, which it also created. When paired with one (or multiple) data storage repositories, such as a database or object store, Presto can function much like a cloud data warehouse.
While Presto can help simplify data operations due to its ability to simultaneously query multiple repositories and combine the results (thereby eliminating the need to build elaborate ETL routines to move data), it’s not a simple piece of software to run. Born in the Hadoop age, there are a lot of knobs to turn, and many organizations struggle to run Presto effectively.
That’s the big reason why a managed service for Presto makes so much sense, says Steven Mih, the CEO and co-founder of Ahana.
“Trying to do Presto yourself has a number of problems,” Mih says. “Everybody knows Presto, but to get it running, not everyone can run it well. The big Internet companies…are the ones who have very experienced database people who can get that to happen. People try to run it in the cloud, and they find it difficult to use. They try to use Athena. But we say, if you want to do Presto, we can get you running in 30 minutes.”
Ahana is offering Presto running as a managed service on AWS. That gives customers the ability to control their Presto environments, which is something that’s not possible with Amazon Athena, AWS’s serverless implementation of Presto.
“It’s all point and click,” Mih says of Ahana. “You don’t even need to know that it’s cloud native underneath. It’s running on EKS [Elastic Kubernetes Service]. It’s running all the cloud native services underneath, and it’s all abstracted.”
Athena, by contrast, runs as a shared service atop Amazon Elastic MapReduce (EMR), which is AWS’s Hadoop-like environment. As a serverless offering, Athena is “non-deterministic,” Mih says, meaning that your performance will vary.
“You could have a query [on Athena] that runs for 6 seconds and then a query that runs for 30 seconds,” he tells Datanami. “Sometimes when there’s not a lot going on with Athena, you get great performance. When there’s stuff going on, you don’t get great performance.”
Ahana’s offering gives users predictable costs. There’s a $5 per day cost to keep the Ahana service running, and then users pay for the Presto services that they consume. In addition to S3, Ahana can query Amazon Relational Database Service (Amazon RDS) for MySQL, Amazon RDS for PostgreSQL Service, and Amazon Elasticsearch. The service also integrates with AWS Glue and Hive Metastores, and it also features its own built-in Hive Metastore (required for Presto) to manage metadata for data stored in S3.
Ahana, which won a 2020 Datanami’s Editor’s Choice Award for Best Big Data Startup, also announced a new collaboration with Intel. The San Mateo, California company is participating in the Intel Disruptor Program, and will offer an Open Data Lake Analytics Accelerator Package for Ahana Cloud that leverages Intel Optane on the cloud with AWS.
“We look forward to working with Ahana and helping bring this compelling open data lake analytic solution to market,” said Arijit Bandyopadhyay, CTO of enterprise analytics & AI within Intel’s data platform group. “As more companies require data to be queried across many different data sources like Amazon S3, Amazon Redshift, and Amazon RDS, Presto will become even more mission critical.”
The companies plan to collaborate in the future to improve how PrestoDB runs on Intel processors and its Optane memory technology. Bandyopadhyay sits on the board of the Presto Foundation, the organization founded to govern the PrestoDB project, along with Mih.
While Ahana is just now announcing GA of its service, it already has a handful of clients, Mih says. O ne of those is the ReferralExchange, a software as a service provider for real estate brokers.
“Within about an hour we were up and running Presto queries on our data, without having to worry about Presto under the covers,” Shiek said in a press release. “With out-of-the-box integrations with a Hive data catalog and no configurations needed, Ahana Cloud takes care of the intricacies of the system, allowing our team to focus on deriving actionable insights on our data.”
PrestoDB is the original version of Presto as it emerged from Facebook. There is a fork of the Presto project called PrestoSQL, which is promoted by Starburst and governed by the Presto Software Foundation, which Ahana is not a part of.