quickstep-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Julian Hyde <jh...@apache.org>
Subject Re: [FINAL DRAFT] Incubator PMC Board Report - June 2016 -- feedback
Date Tue, 07 Jun 2016 02:01:52 GMT
For those of you not on general@incubator, forwarding Marvin Humphrey’s comments on Quickstep’s
report.

(I recommend that PPMC members join general@incubator… you will learn a lot about incubation
& the Apache way.)

Julian


> On Jun 6, 2016, at 6:40 PM, Marvin Humphrey <marvin@rectangular.com> wrote:
> 
>> --------------------
>> Quickstep
>> 
>> Modern servers pack enough storage and computing power that just a decade
>> ago was spread across a modest-sized cluster. Given that we are on a
>> technological path to continue to increase the storage and compute densities
>> of individual server nodes, we must complement methods that focus on
>> '''scaling-out''' by also developing methods to '''scale-in''' to fully
>> exploit the hardware capabilities that is packed in each server node. The
>> initial phase of the Quickstep project focuses on this scaling-in aspect.
>> Quickstep uses novel methods for organizing data (including columnar and
>> hybrid storage organization), template metaprogramming for vectorized query
>> execution, and a query execution paradigm that separate control-flow from
>> data-flow. Collectively, these methods achieve high performance on
>> contemporary servers with multi-socket, multi-core processors and large main
>> memory configurations. To keep the project focused, the project’s initial
>> target is interactive in-memory data warehousing workloads in single-node
>> settings. In the near future we plan to expand from this initial single-node
>> focus to a distributed setting. Early results indicate that Quickstep is
>> over an order-of-magnitude faster than existing platforms including Spark
>> 2.0 and PostgreSQL 9.6 Beta1 (that now has parallel query processing).
> 
> The description of Quickstep is longer than it needs to be and shouldn't
> contain wiki markup.  Please edit podlings.xml.
> 
>> Three most important issues to address in the move towards graduation:
>> 
>>  1. Acquire early adopters
>>  2. Acquire early adopters
>>  3. Acquire early adopters
>>    (We know it is that important!)
> 
>>  Roman Shaposhnik (rvs):
>> 
>>    Really, at least goal #2 should be "make a first ASF release". Hopefully
>>    we can work with the community to get a sense how important it is in
>>    addition to growing a community (all 3 goals in this report).
> 
> +1 to this excellent comment of Roman's on the Quickstep report.  "Release
> early, release often" is a technique for acquiring early adopters!



Mime
View raw message