NoSQL panel for OSCON 2010
David Strauss
they typically involved each project differentiating itself from
SQL-style systems. While that is helpful to many people, it's largely
redundant from presentation to presentation. It also doesn't exploit
OSCON's primary benefit: bringing together many FOSS projects.
Instead of (or maybe in addition to?) the separate presentations, I
would like to assemble a panel of representatives from major NoSQL and
related projects to present at OSCON. While a brief introduction would
define how NoSQL differs from traditional relational systems, the
primary focus would be differentiating NoSQL projects from each other.
Because these projects tend to focus on the "common enemy" of RDBMS, I
believe this area is under-explored.
Some potential classes of differentiation:
* Sharding and partitioning
* Replication, consistency, and managing conflict
* Query and view support
* Validation/schema support
* Language support (PHP, C++, etc.)
* Ideal use cases
I'm Drupal's MySQL maintainer and one of the engineers of Drupal's Field
API, which supports modular backends with a focus on storage in
NoSQL-style systems. I would be happy to moderate and help coordinate
the topics of discussion and debate.
If I can get commitment from some major project representatives, I will
also spearhead the OSCON presentation proposal.
--
David Strauss
| da...@fourkitchens.com
| +1 512 577 5827 [mobile]
Four Kitchens
| http://fourkitchens.com
| +1 512 454 6659 [office]
| +1 512 870 8453 [direct]
Doug Judd
I'm the original creator of Hypertable and the CEO of Hypertable, Inc. I would be happy to participate in a NOSQL panel at OSCON 2010 and can represent the Bigtable class of NoSQL technologies.
- Doug
Emil Eifrém
I'm a co-founder of the Neo4j graph database project (presented at
OSCON 2009). I'd be happy to represent graph databases in a panel at
OSCON 2010.
Overall, I think too few presentations these days zoom out and put the
various NOSQL projects in perspective.[1] Which I think is what more
people need at this stage rather than low-level details about specific
projects. Hopefully a well-prepared, well-run panel could provide some
level of that.
1] I added my take on the space to my nosql east presentation
(http://www.slideshare.net/emileifrem/nosql-east-a-nosql-overview-and-the-benefits-of-graph-databases)
and in writing (http://blogs.neotechnology.com/emil/2009/11/nosql-scaling-to-size-and-scaling-to-complexity.html).
Anyway, would be happy to participate.
Cheers,
-EE
> --
>
> You received this message because you are subscribed to the Google Groups
> "NOSQL" group.
> To post to this group, send email to nosql-di...@googlegroups.com.
> To unsubscribe from this group, send email to
> nosql-discussi...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/nosql-discussion?hl=en.
>
--
Emil Eifrem [em...@eifrem.com]
http://blogs.neotechnology.com/emil
http://twitter.com/emileifrem
http://eifrem.com
Jonathan Ellis
what I modestly believe to be the best introduction to the NoSQL
space: http://www.rackspacecloud.com/blog/2009/11/09/nosql-ecosystem/.
I'd be happy to participate.
-Jonathan
Justin Sheehy
On Tue, Dec 22, 2009 at 4:54 PM, David Strauss <da...@fourkitchens.com> wrote:
> Instead of (or maybe in addition to?) the separate presentations, I
> would like to assemble a panel of representatives from major NoSQL and
> related projects to present at OSCON.
I think that this is a great idea.
> If I can get commitment from some major project representatives, I will
> also spearhead the OSCON presentation proposal.
I am happy to participate, and can represent Riak.
-Justin
J. Chris Anderson
On Dec 22, 3:54 pm, David Strauss <da...@fourkitchens.com> wrote:
> While OSCON 2009 had several separate presentations on NoSQL projects,
> they typically involved each project differentiating itself from
> SQL-style systems. While that is helpful to many people, it's largely
> redundant from presentation to presentation. It also doesn't exploit
> OSCON's primary benefit: bringing together many FOSS projects.
>
> Instead of (or maybe in addition to?) the separate presentations, I
> would like to assemble a panel of representatives from major NoSQL and
> related projects to present at OSCON. While a brief introduction would
> define how NoSQL differs from traditional relational systems, the
> primary focus would be differentiating NoSQL projects from each other.
> Because these projects tend to focus on the "common enemy" of RDBMS, I
> believe this area is under-explored.
>
> Some potential classes of differentiation:
> * Sharding and partitioning
> * Replication, consistency, and managing conflict
> * Query and view support
> * Validation/schema support
> * Language support (PHP, C++, etc.)
> * Ideal use cases
I'm a CouchDB committer and I'd be happy to represent CouchDB.
Thanks for the idea and the energy to put this together.
Chris
ryanobjc
project and working on bigtable-alike systems for years now, he is a
natural choice to represent one of the many dynamic and successful
"nosql" (i hate that term) systems.
Go Stack!
-ryan
K.S. Bhaskar
and will gladly participate on a panel.
Regards
-- Bhaskar
ks dot bhaskar at fnis dot com <-- send e-mail here
> signature.asc
> < 1KViewDownload
David Strauss
our Google Group, I've created a page to coordinate here:
https://wiki.fourkitchens.com/display/PUBLIC/NoSQL+OSCON+Panel+Proposal
David Strauss
below in case anyone has last-minute changes (which must be in by
midnight today, where midnight is probably for UTC-8). I don't mind even
getting a call on my mobile (listed in my sig) if you have an urgent
change or would like to discuss things real-time.
About one-third the people did not have existing bios on the OSCON site,
so I used existing bios found by Meghan Hill (of 10gen) or wrote brief
ones myself that highlighted relevance to this panel and mentioned their
companies. Please review your bio (on your OSCON account) and add any
additional information you would like.
I did not intentionally order the speakers to put myself on top; I just
entered myself first. OSCON's reordering tool seems very broken. I have
no idea what order speakers will show up in for conference materials.
== Title ==
Non-Relational Rebellion: The NoSQL Project Panel
== Description ==
Meet the rebels against SQL and relational storage engines. Following a
brief introduction to "NoSQL" systems, core developers from CouchDB,
MongoDB, Riak, Cassandra, Neo4j, and Hypertable will differentiate their
projects and make the case for non-relational storage running your next
project. The panel will be moderated by the MySQL maintainer of the
Drupal project.
== Topics ==
Cloud Computing, Databases, Trends, Java, PHP, Tools & Techniques,
Python, Ruby
== Abstract ==
Brief introduction presented by Tim Anglade:
Brief introduction presented by Tim Anglade:
# *40 years in the desert*: yes, the relational model is actually
celebrating its "40th
birthday":http://portal.acm.org/citation.cfm?id=362685 this year. Let's
take an unbiased look about what’s great and not-so-great about it.
# *A not-so-novel idea*: precursors & roots of NoSQL systems from the
70s to the 00s. NoSQL projects are not just defined in opposition to
SQL; they come from approaches as old or sometimes older than the
relational model itself.
# *The odd -couple- family*: why those very separate projects, efforts
and models have been grouped together under one brand in the public’s
mind. Plus, a possible categorization of the sub-families within the
movement.
# *“What's in a name? That which we call a NoSQL system, by any other
name would be as frackin’ sweet”*: Obligatory, rapid mention of the
controversy around the name and why nobody should care.
# *SQL vs. NoSQL, Live from… THE THUNDERDOME*: a breakdown of where
NoSQL sub-families tend to shine and where SQL should probably be used
instead.
An initial set of prepared panel questions will focus on:
* Approaches to replication and scalability
* Example ideal uses
* Language and platform integration capability, maturity, and performance
* Whether we should expect a shakedown in projects or continued plurality
== Level ==
Intermediate
== Additional Notes ==
We plan to reserve at least 10 minutes for audience questions, but
scheduling the panel directly before lunch would help support the
further discussion that will likely follow.
== Tags ==
nosql, couchdb, mongodb, hbase, riak, cassandra, hypertable, sql,
sharding, replication, json, bson
== Speakers ==
David Strauss da...@fourkitchens.com Four Kitchens
J Chris Anderson jch...@apache.org couch.io
Justin Sheehy justin...@gmail.com Riak/Basho
Jonathan Ellis jbe...@gmail.com Apache Cassandra/Rackspace
Emil Eifrem em...@eifrem.com Neo Technology
Doug Judd nugge...@gmail.com Hypertable
Dwight Merriman dwi...@10gen.com 10gen
Tim Anglade tim.a...@gmail.com Tin
Emil Eifrém
like the problem with panels are always that you don't have time to
get a coherent message out. With intro + quite a few people on the
panel, we may run short. Is it scheduled for 40 or 60 mins or more?
Also, would love it if neo4j was part of the tags. :)
-EE
--
David Strauss
> Okay everybody, I've submitted the panel to OSCON. I've put the proposal
> below in case anyone has last-minute changes (which must be in by
> midnight today, where midnight is probably for UTC-8). I don't mind even
> getting a call on my mobile (listed in my sig) if you have an urgent
> change or would like to discuss things real-time.
Just to clarify, it will be final at midnight, UTC-8. That's just over
eight hours from now.
Emil Eifrém
> On 2010-01-31 23:32, David Strauss wrote:
>> Okay everybody, I've submitted the panel to OSCON. I've put the proposal
>> below in case anyone has last-minute changes (which must be in by
>> midnight today, where midnight is probably for UTC-8). I don't mind even
>> getting a call on my mobile (listed in my sig) if you have an urgent
>> change or would like to discuss things real-time.
>
> Just to clarify, it will be final at midnight, UTC-8. That's just over
> eight hours from now.
I think it's actually ~32h from now. 11:59pm PST Feb 1st.
Cheers,
David Strauss
> Looks good overall. Only concern I'd have is that of time. It seems
> like the problem with panels are always that you don't have time to
> get a coherent message out. With intro + quite a few people on the
> panel, we may run short. Is it scheduled for 40 or 60 mins or more?
It's scheduled for 40 minutes, but I've requested scheduling just before
lunch so we can continue discussions and not feel bad about either going
over or leaving little time for audience questions.
There is no option to make the session longer without making it a
three-hour tutorial.
> Also, would love it if neo4j was part of the tags. :)
Done. (And sorry for the inadvertent omission.)
Krishna Sankar
Cheers
<k/>
David Strauss
> I have proposed a tutorial "Building a semi-structured Data Cloud using NOSQL Datastores". If selected, we could continue some discussions there as well. Am not sure how the timings would be.
Tutorials happen in the days before sessions.
David Strauss
> below in case anyone has last-minute changes (which must be in by
> midnight today, where midnight is probably for UTC-8). I don't mind even
> getting a call on my mobile (listed in my sig) if you have an urgent
> change or would like to discuss things real-time.
Should I make any last-ditch effort to include people from Redis and
Tokyo Cabinet/Tyrant? If not, we should at least include them in any
overviews.
Tim Anglade
> On 2010-01-31 23:32, David Strauss wrote:
>> Okay everybody, I've submitted the panel to OSCON. I've put the proposal
>> below in case anyone has last-minute changes (which must be in by
>> midnight today, where midnight is probably for UTC-8). I don't mind even
>> getting a call on my mobile (listed in my sig) if you have an urgent
>> change or would like to discuss things real-time.
>
> Should I make any last-ditch effort to include people from Redis and
> Tokyo Cabinet/Tyrant? If not, we should at least include them in any
> overviews.
I think it'd be nice to have them in, as they are fairly prominent
projects. On the other hand, the panel might be too crowded already.
Not much fun if every participant can only utter one sentence during
the session. If too many people jump in, maybe it'd make sense to have
one representant by “type of architecture” (document, key-value,
graph, etc.) instead?
As far as an overview is concerned, they're definitely on my list of
projects and architectures covered.
(For people who don’t know: I submitted a proposal for a NoSQL
overview talk. Ideally it would precede the panel to make it a nice
double-header of intro + advanced discussion.) Dunno if anybody else
submitted another NoSQL overview proposal.
Cheers,
Tim
Sergio Bossa
> As far as an overview is concerned, they're definitely on my list of
> projects and architectures covered.
Tim,
you may want to take a look at Terrastore too, which is a (rather new)
distributed *and* consistent document store:
http://code.google.com/p/terrastore
Feel free to ask if you need any additional information.
It would be great to have it on your overview.
Cheers!
Sergio B.
--
Sergio Bossa
Software Passionate and Open Source Enthusiast.
URL: http://www.linkedin.com/in/sergiob
Tim Anglade
> On Mon, Feb 1, 2010 at 1:54 PM, Tim Anglade <tim.a...@gmail.com> wrote:
>
>> As far as an overview is concerned, they're definitely on my list of
>> projects and architectures covered.
>
> Tim,
>
> you may want to take a look at Terrastore too, which is a (rather new)
> distributed *and* consistent document store:
> http://code.google.com/p/terrastore
> Feel free to ask if you need any additional information.
> It would be great to have it on your overview.
>
> Cheers!
>
> Sergio B.
Thanks Sergio, Terrastore was already on my list.
Here's what I have so far:
- CouchDB
- MongoDB
- Riak
- Redis
- TokyoCabinet
- Neo4J
- InfoGrid
- Sones
- HyperGraphDB
- Hypertable
- SimpleDB
- Terrastore
- Hadoop
- Mnesia
- Cassandra
- HBase
- JackRabbit
- Voldemort
- GT.M
- Dynomite
- MemcacheDB
- BigTable
- Dynamo
- Sherpa
- Oracle's Spatial Layer (non-free)
- ESRI ArcGIS (non-free)
- SAND (academic Geographical/Spatial database)
Obviously, that's quite a lot to cover so I don't plan on going
in-depth on each project. I think there are enough project-specific
presentations and online ressources to really learn about each of them
so I'll stick to trying to present the big picture, subfamilies,
origins of the movement, issues, future, etc.
Cheers,
Tim
Sergio Bossa
> Thanks Sergio, Terrastore was already on my list.
That's great, thanks much.
> Here's what I have so far:
> [CUT]
> Obviously, that's quite a lot to cover so I don't plan on going
> in-depth on each project. I think there are enough project-specific
> presentations and online ressources to really learn about each of them
> so I'll stick to trying to present the big picture, subfamilies,
> origins of the movement, issues, future, etc.
Yep, that's a really huge list.
Anyways, feel free to ask if you need any particular information about
Terrastore or how it compares to other solutions.
김형준
http://www.slideshare.net/babokim/neptune-distributed-data-system
http://www.openneptune.com/
Feel free to ask if you need any additional information.
Thanks.
hjkim.
2010/2/1 Tim Anglade <tim.a...@gmail.com>:
> --
> You received this message because you are subscribed to the Google Groups "NOSQL" group.
> To post to this group, send email to nosql-di...@googlegroups.com.
> To unsubscribe from this group, send email to nosql-discussi...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/nosql-discussion?hl=en.
>
>
--
오픈소스 대용량DB
http://dev.naver.com/projects/neptune
Steven Noels
There is other bigtable clone open source.
http://www.slideshare.net/babokim/neptune-distributed-data-system
http://www.openneptune.com/
Feel free to ask if you need any additional information.
Interesting. Just to satisfy my curiosity: what are the differences/similarities between Neptune and HBase? The approach and architecture seems quite identical, though I believe Neptune adds a higher-level abstraction. Is that a correct interpretation?
Thanks,
Steven.
Steven Noels http://outerthought.org/
Outerthought Open Source Java & XML
stevenn at outerthought.org Makers of the Daisy CMS
Jean-Daniel Cryans
will so I can represent HBase, I've been a committer since almost 2
years.
J-D
김형준
So we decided own bigtable clone data system.
Difference
1. No appending operation in Hadoop. so HBase's change log isn't safe.
Otherwise Neptune has change log system. Neptune's change log system
supports append.
2. Loosely coupled hadoop
If you implement NeptuneFileSystem interface. you can replace file system.
We already deployed several systems.
searcus.com(twitter search service) stores more than 2TB using
Neptune(10 neptune servers).
Other site has 3 hundred million html pages in Neptune.
Thanks,
hjkim.
2010/2/1 Steven Noels <ste...@outerthought.org>:
Jean-Daniel Cryans
> 1. No appending operation in Hadoop. so HBase's change log isn't safe.
> Otherwise Neptune has change log system. Neptune's change log system
> supports append.
Just to correct a common misconception, appending operation is
supported in Hadoop see
http://hadoop.apache.org/common/docs/r0.20.1/api/org/apache/hadoop/fs/FileSystem.html
The problem for 0.20 was that there was no way to make sure that
appends webt on the wire after the call was issued. Closed
write-ahead-log files were really persisted to disk so we set the max
file size to 64MB to limit loss in case of machine failure.
Fortunately this is all fixed now by HDFS-265 for Hadoop 0.21
J-D