Bridging the gap between open source databases and database services

It is relatively easy for a group of people to create a new database management system or a new data store . The reason why we know this , Because in the past 5 In the calculation of the year , The proliferation of tools that provide structure for data has increased , And it seems to be growing faster and faster . To a large extent, thanks to the innovation of large-scale enterprises, cloud builders and scholars , They just like to wander around the database to prove a point .

however , Turn an open source database or data storage project into a business that can provide enterprise level matching and complete and support a wider range of use cases and customer types and sizes , This is quite another matter . It's hard work , It takes a lot of people 、 attention 、 Money and luck .

This is a Dipti Borkar、Steven Mih and David Simmen stay Launched two years ago Ahana In order to Facebook Created Presto Distributed SQL Engine PrestoDB The task of commercializing variants , It is no coincidence that , With the original Presto The creator of used 了 PrestoSQL, Now known as Trinio, This is a Presto A variation of the , Commercialized by their company , be called Starburst. In any case , these Presto Variations are federated with databases and datastores , And provide a general SQL layer , Allow them to query in place —— This is a very powerful function , This is necessary due to the persistence of legacy databases and the importance of data .

It's too difficult to move them all to one place to query it , This is how companies try to create data warehouses . even so , Data warehouses usually only have summary data , Although once the data enters the warehouse, it has the advantage of convenience , But getting data in the warehouse （ And make sure it's not garbage ） It's a very painful thing . In short , As we said a few months ago , You want to do data analysis without a data warehouse , This is the darling of the database industry Snowflake Using its cloud data warehouse does the exact opposite .

More and more companies want to use PrestoDB To query the location of the data . That's why Ahana Can extend its last year 8 Announced in A Reasons for round financing ,Google Ventures、Lux Capital、Third Point Ventures and Leslie Capital In which... Was raised 2720 Thousands of dollars , In order to increase Ahana Raised 480 Million dollars of seed funds to obtain in 2020 Year begins . With A Extension of the series ,Liberty Global Ventures It is the venture capital Department of the telecom company , More involvement by the same company doing business across Europe and Google ventures , Re funding the series 720 Ten thousand dollars a kitten .（ We strongly suspect Liberty Global yes Ahana The customer , But the CEO Steven Mih Will not comment on this .） This brings the total revenue so far to 3200 Thousands of dollars ,Mih Add ,Ahana Not going to raise money . In the current economic environment , We joked that , If someone gives you money , You will find a reason to accept it .

In the first round A In the ten months of round financing ,Ahana The number of employees has more than doubled , Less than 50 people , And has downloaded more than 100,000 Copy its Ahana Implemented PrestoDB copy .Mih It is impossible to say that it is at the commercial level of the database Ahana Cloud How many paying customers are there in the implementation .

As for increasing the company's payroll ,Mih Your caution is understandable .“ We want to know what is happening in the global economy and the possible adverse factors associated with it ,”Mih tell The Next Platform, Avoid using R word .“ If some potential problems do not occur , Then our development speed will be very fast .”

This growth is driven by the need for joint queries across database platforms , The concept of multimodal data processing makes this even more obvious ,Matt Bornstein、Jennifer Li and Martin Casado（ One of them OpenFlow The founder and Nicira One of the co founders of ,Nicira by VMware Provides NSX Virtual network stack ）, All these people are working around the world for Andreesen Horowitz Make good technology investment .

The core of this modern data processing architecture is the so-called data lake —— Part of the data warehouse comes from the past , Part of the data Lake comes from Hadoop Time , But it's really just cheap and deep storage , No need to use MapReduce To solve the problem of unstructured data on machine clusters .

This is from Mih The chart of the chart summarizes the center of the chart more clearly and easily ：

“ As you know , A lot of data is injected into the data lake , Including semi-structured 、 Structured and unstructured data ,”Mih explains .“ As everything is commercialized , People will ask why they should put data in another proprietary store , For example, data warehouse , And why they should keep it in an open format . If they do try to put this data into commercial storage , So computing on the data warehouse is proprietary . The idea of the data lake library is to use open source computing , And for SQL Query processing Presto Is one of the main options . And then for non SQL Queries and workloads , You can use ML and AI Frame calculation , And use Parquet Equiform . Storage is the commodity of the lakeside cottage , The computing layer is really where the cost lies ,

The entire multimodal data processing architecture has many mobile parts , If Ahana To succeed in PrestoDB Commercialize and federate distributed query engines in various relational data stores , It must become easier to install and test data Lake libraries SQL The core . This is the new Ahana Cloud for Presto The full meaning of the Community Edition . It is a free and unrestricted database version , Can run on any single cluster , Regardless of the size .（ majority Presto Customers have multiple clusters , This is where the subscription will begin .） The following is the community version and the complete Ahana Cloud for Presto The difference between versions ：

Community Edition stay Amazon Web Services Run on the cloud , It's like Presto The production of Ahana Cloud equally , As long as it only runs on a single cluster —— No matter how many EC2 Instance drives it ——Community Edition It's all free . There are some warnings . The Community Edition doesn't support Graviton、Graviton2 or Graviton3 example , It has only community support . If you want Ahana Cloud for Presto Enterprise version , You can seamlessly upgrade to it , Then you can have any number of clusters and run them in any AWS Run on instance type , Include AWS Created for it Graviton Arm The server CPU Series for your own use . The production version also has higher security 、 Performance enhancements （ for example AWS Autoscale on ）, Of course, Ahana Technical support provided by hired real people .

Now? ,Ahana It allows people to quickly start using Presto, And save them setting up on the data lake Presto It takes days or weeks . Just grab this container , To open it , Point it to the data Lake , Then start using SQL Query to attack it . every last Community Edition Users can use it permanently , And before you have a second cluster or need enhanced security or performance , Never pay for it .

author ：Timothy Prickett Morgan

Source of the article ：https://www.nextplatform.com/2022/06/17/bridging-the-gap-between-open-source-database-and-database-business/