DocumentDB

I haven't delved into the MSFT pricing model, but DynamoDB is pay-for-through put. You provision your table with a certain amount of read/write performance that is guaranteed, and you pay for that.

cvburgess4• 12 years ago

I'd assuming pricing will be similar and incredibly competitive... and thanks, didn't realize how limited dynamoDB was

ryanfitz• 12 years ago

I've been using DynamoDB since it was released. I wrote a nodejs driver[1] and a nodejs data mapper for it[2], so I have a decent bit of experience with it. Browsing the DocumentDB docs the two services seem to be very different. DynamoDB is really just a key value store with some very nice properties, but also a lot of tradeoffs. One such tradeoff is querying data is very limited in DynamoDB. In Dynamo you can only query data by its primary hash key and optional range key. These keys you must specify upfront when you create your table and cannot be changed afterwards.

DocumentDB seems much more similar to mongodb and appears to have a very flexible query ability. In my opinion, one of the best features with DynamoDB is you can tune the number of reads/writes each individual table requires. This lets you scale up and down your database and greatly helps keep costs down. This is a feature that only a hosted database service can offer. I haven't yet read any pricing info on DocumentDB, but hopefully they offer a similar feature as this is really where a hosted database service can shine.

[1]https://github.com/Wantworthy/dynode [2]https://github.com/ryanfitz/vogels

cvburgess4• 12 years ago

Thanks, yeah I can see how this is more like hosted mongoDB - and I've seen dynode on npm, nice work!

pokstad• 12 years ago

Sounds very similar to CouchDB. Server side Javascript written by the user, and an HTTP interface. The ability to adjust consistency is really neat.

Goranek• 12 years ago

similar to cloudantdb a little bit more because of searching and indexing.

andrewguenther• 12 years ago

The ability to adjust consistency isn't really a new concept. The first I saw it was in Amazon's DynamoDB paper in 2007. Cassandra uses the model outlined there as well, and I'm sure there are others out there that I'm less familiar with.

ahoge3• 12 years ago

CouchDB doesn't support ad hoc queries. CouchDB is all about MapReduce and heavy caching. It's very rigid.

bkeroack• 12 years ago

...and MS goes after MongoDB. It would be nice to see an on-premises version, if only to compare performance/consistency with Mongo.

codereflection• 12 years ago

Not really so far, it's Azure only.

taude• 12 years ago

I have to believe it's coming, since a lot of MSFTs core business is still selling server-licenses for business who host their own data centers. I for one would love to use something like this, and our company is one that's not comfortable with third-party open-source tools, and we have to host in our own data center...

reubenbond7• 12 years ago

DocDB is built atop a very well battle-tested distributed systems framework with replication based on a multi-Paxos implementation. I haven't done any benchmarks, but the replication model is far superior to MongoDB's "journal journal" model for performance. http://daprlabs.com/blog/blog/2014/08/22/azure-documentdb/

dougws• 12 years ago

Your objections to MongoDB's model seem reasonable, but I don't see any evidence in either this comment or the linked blog post that DocumentDB is better (especially in the absence of benchmarks). What is this "battle-tested distributed systems framework"? Several of your complaints about MongoDB have to do with the interaction between persistence to disk and replication to the network; as the Multi-Paxos algorithm does not specify when data should be written to disk (much less what the format should be), what reason is there to believe that DocumentDB does this any better?

I'm totally willing to believe that DocumentDB beats the pants of MongoDB on just about every axis (in fact, that seems pretty likely) but it's going to take some actual numbers and a better description of the internals.

hendzen• 12 years ago

Another way to look at this is that it is based on Microsoft's internal distributed systems infrastructure and thus will never be open-sourced for the same reason Google will never open-source Spanner, Megastore or Colossus. Having battle-tested internal systems to build on is nice, but it means the DocumentDB code is probably nearly impossible to run outside of Microsoft.

theatraine• 12 years ago

I wouldn't be surprised if it happens, Microsoft has on-premise versions of several Azure offerings (i.e. WebSites, IaaS, Service Bus) which are released in Windows Azure Pack (http://www.microsoft.com/en-us/server-cloud/products/windows...). It might take some time though.

cbsmith4• 12 years ago

Honestly, this seems a bit more like Cassandra or Couchbase than Mongo.

jnardiello2• 12 years ago

And that is a creative product name. Well played MS.

talles3• 12 years ago

Just like SQL Server

programminggeek• 12 years ago

It's really smart. It's a document database, so they call it DocumentDB. It's better than calling it something like Dolphin and then explaining that it's a document database. Simple and clear is a good thing.

allegory5• 12 years ago

Err, Apple:

Airport, Bootcamp, Safari, Time Machine, FaceTime, Grand Central Dispatch, QuickTime...

and I won't even start on the open source side of things...

talles3• 12 years ago

But it's also kinda selfish don't you think?

tomsthumb• 12 years ago

"Microsoft has DocumentDB service now."

"Oh cool, which document database is it? MongoDB? CouchDB? Cassandra?"

"It's DocumentDB!"

"...."

tdicola• 12 years ago

Wouldn't have surprised me to see it called something like Microsoft NoSQL Database Cloud Edition for Agile Developers Powered By JavaScript 2014.

jpalomaki• 12 years ago

Ad hoc queries using SQL like syntax. No need to define indexes.

Javascript execution within database. Stored procedures, triggers and functions can be written with Javascript. "All JavaScript logic is executed within an ambient ACID transaction with snapshot isolation. During the course of its execution, if the JavaScript throws an exception, then the entire transaction is aborted."

Pricing is based on "capacity units". Starts with $22.50 per month (this includes 50% preview period discount). One capacity unit (CU) gives 10GB of storage and can perform 2000 reads per second, 500 insert/replace/delete, 1000 simple queries returning one doc.

In order to see pricing details, change the region to "US West": http://azure.microsoft.com/en-us/pricing/details/documentdb/

Very interesting addition to Microsoft offering. I was actually just yesterday wondering if they have any plans for this kind of service. Table Storage is quite primitive and Azure SQL on the other hand gets expensive when you have lots of data.

One potential "problem" with this is the bundling of storage capacity and processing power. If I understand this correctly, I would need to buy 10 CUs per month to store 100GB of data even if I'm not very actively using that data.

jnardiello2• 12 years ago

That is actually good to start with, yet scaling will cost A LOT.

talles3• 12 years ago

Can I use DocumentDB out of Azure (hook my own)?

twotwotwo• 12 years ago

You know, it's not a simple question, but there's some case for Microsoft just up and releasing some scale-out DB for free (as in beer) or at least cheap.

My thinking is, scale-out deployments aren't all that likely to pay at SQL-Server-like rates per CPU anyway, and you're helping give the Windows world better parity with the horde of Linux options for folks that need scale-out (or, just as commonly, hope to need it someday).

On the flipside, by doing that you might forego some sales of SQL Server (though I suspect that's limited; most folks that need SQL Server really need SQL Server) or sales on Azure that could theoretically help recover the dev costs. But greatly improving the scale-out-on-Windows story seems like a big deal, the kind of thing that might justify going to lots of effort to make a DB product then giving it away.

aikah• 12 years ago

You know, it's not a simple question

Yes it is a simple question, and the answer is NO.

ceejayoz3• 12 years ago

http://azure.microsoft.com/en-us/documentation/articles/docu...

Want to edit or suggest changes to this content? You can edit and submit changes to this article using GitHub.

Pretty remarkable given Microsoft's approach to open source in the 1990s that they're now using a service built around Linus's bespoke open source version control system to allow people to suggest changes to their documentation.

pjmlp2• 12 years ago

Microsoft is the new IBM. :)

braum• 12 years ago

Agree and I've been saying it for ten damn years, havn't I been saying it!

scholia2• 12 years ago

Nothing remotely like it, in terms of power, range, degree of market control, or willingness to be commercially evil. IBM used to be more than twice as big as every other IT company put together, and in some years, made more than 100% of the industry's profit.

Even today, after disposing of numerous divisions, it's still a $100bn company, ie still larger than Microsoft. (So you pay IBM more than you pay MS every year, and you always have done.)

However, it is true that Microsoft is the IBM of PC software, just as Google is the IBM of the web, Intel is the IBM of processors, Facebook is the IBM of social networking, Cisco is the IBM of routers, Oracle is trying to be the IBM of server software, Apple is the IBM of hipster status symbols... Well, you get the idea.

IBM used to be the IBM of everything in IT, and started getting into other areas (telephone switches, copiers, cash machines etc) before the wheels started to come off.

Today, Google is a lot more powerful than Microsoft, much scarier, and much more ambitious in IBM-like ways (self-driving cars, robots, superbrains etc).

rbanffy• 12 years ago

IBM still does cool things that are completely beyond Microsoft's grasp. The breadth of their research is still very impressive - from basic physics all the way to some of the most functional AI around. When was last time Microsoft did a product that could beat a human on Jeopardy?

If they want to be the next IBM, they have a lot of work ahead of them.

allegory5• 12 years ago

Beware a wolf in sheep's clothing...

Edit: the downvoters clearly weren't in the industry in the 80's and 90's and haven't dealt with them in the enterprise / volume licensing department recently. Comparing the two, they're even more ruthless, unfair and incompetent than ever and will screw you as hard as they can once you're locked in.

This stuff gets you through the door, as does BizSpark etc, then you're not a friend but a cow for the milking via VL, audits and licensing changes.

I speak from experience working for 4 paid up gold partners over the last decade and then dealing with them in a corporate capacity back to '95. Every game ends the same.

Edit 2: appears you can't tell the truth about Microsoft these days in the same fashion you couldn't tell the truth about Apple about 3-4 years ago...

chipotle_coyote• 12 years ago

Either that or the downvoters are willing to consider the possibility that 25 years -- not to mention a radically different market position -- makes a difference. The Microsoft of 2014 is not the Microsoft of 1989.

...in the same fashion you couldn't tell the truth about Apple about 3-4 years ago...

Yeah, if only there had been lots of people highly critical of Apple in 2011. Whatever happened to that Android thing, anyway?

JeremyMorgan• 12 years ago

I've been trying to tell everyone this in the last year. Things have really changed in the Microsoft world. For the better.

It makes me forget they killed my pappy!

mythz5• 12 years ago

Because they've released a vendor lock-in Azure-only, proprietary NoSQL database, 5+ years after everyone else?

tacticus• 12 years ago

So what about their lawyers campaigning for the copyrighting of APIs?

mythz5• 12 years ago

Where does it say DocumentDB is Open Source? The lack of any mention is usually a good indicator that it's not.

dkuntz2• 12 years ago

That comment was about Microsoft's documentation...

ceejayoz3• 12 years ago

Nowhere does it say DocumentDB is open source.

The Microsoft of the 1990s wouldn't be using Github.

datashaman• 12 years ago

A comment from someone on the team further up the page mentions that ElasticSearch is 'leveraged' inside the service, so I'd say there should be some mention of the Apache 2 license to represent the use of ElasticSearch.

https://github.com/elasticsearch/elasticsearch

bigtones2• 12 years ago

Dude, that's just the article, not the source. Eg: Github as CMS.

ceejayoz3• 12 years ago

I'm fully aware of that, considering I quoted the bit talking about contributing to the article.

andrea_s5• 12 years ago

Am I alone in thinking that sql-like syntax is actually a step backwards from building query documents programmatically (MongoDB style)?

haxor1232123• 12 years ago

yes

balls1872• 12 years ago

Not necessarily.

I would say it allows flexibility in thinking about your storage and access patterns, which in turn allows flexibility in use. Hopefully without the expense of performance, and integrity.

luuio4• 12 years ago

how? you're just doing the same thing with MongoDB in their own json-like syntax instead of the T-SQL that everyone knows

andrea_s5• 12 years ago

For simple queries there is virtually no difference (and clearly one should favor what he knows best).

But there are cases where you need to build the query object programmatically (and dynamically), and I believe it's a bit awkward to do that to obtain an SQL statement. But possibly it's just my personal taste :)

reubenbond7• 12 years ago

You can still build documents programmatically. The SQL-esque syntax is optional. The primary interface is HTTP.

andrea_s5• 12 years ago

Oh, I see, I've missed this detail on the first read. Thanks!

ahoge3• 12 years ago

The SQL-esque syntax is optional. The primary interface is HTTP.

Then the introduction needs some work.

"DocumentDB enables complex ad hoc queries using the SQL dialect"

"Azure DocumentDB offers the following key capabilities and benefits: Ad hoc queries with familiar SQL syntax"

The "Query DocumentDB" article also seems to be focused on that SQL dialect.

johnnymonster• 12 years ago

I've been a long time elasticsearch user and can say that I personally HATE sql syntax when it comes to querying data in mysql, just because I usually can't get the joins right etc to build a search that is reasonable. However, they might be taking this approach because sql syntax is familiar. JSON data structures for querying over HTTP was used with elasticsearch because they were familiar to most developers. I think overall it helps adoption if you have layers that are not completely foreign to the implementing developers. They could have written their query syntax in protocol buffers or coffescript or what ever the hell they wanted really, but I think that would have hurt adoption.

ahoge3• 12 years ago

I'd say RethinkDB does it best. You just chain some regular function calls and pass some lambdas here and there. It looks and feels very natural. Plus, that "data explorer" thing lets you test those queries quickly.

Generating a list of all used tags:

  r.table('posts')('tags').reduce((a, b) => a.setUnion(b))

Grabbing all posts with the 'foo' tag:

  r.table('posts').filter((doc) => doc('tags').contains('foo'))

reubenbond7• 12 years ago

DocumentDB has a similar looking (maybe not as powerful) interface:

  context.collection.filter(function(doc) {
    return doc.city == "Melbourne" && doc.rating > 4.5;
  }).map(function(doc) {
    return doc.displayName;
  });

streptomycin• 12 years ago

DocumentDB utilizes a highly concurrent, lock free, log structured indexing technology to automatically index all document content. This enables rich real-time queries without the need to specify schema hints, secondary indexes or views.

How does that work? Isn't that going to incur a major performance hit? If not, why don't other databases get rid of indexes?

Also, if anyone from MS is reading, http://azure.microsoft.com/en-us/documentation/articles/docu... links to http://azure.microsoft.com/en-us/documentation/articles/docu... which is a 404 error.

andrea_s5• 12 years ago

One way to approach a behind the curtain automatic index generation would be to gather data about collection usage, and use such data to enable targeted indexes.

I don't think it would be feasible to have "indexes for everything", as the numbers involved would scale geometrically.

arthurcolle• 12 years ago

Forgive my ignorance but why would it scale geometrically? I tried to quickly think of a complexity analysis and I don't see where you are coming from.

utunga• 12 years ago

Another case of Not Invented Here syndrome from Microsoft. One wonders why they couldn't just take the open source and very well architected RavenDB http://ravendb.net .Net Document DB and provide first class support for that within Azure.

Maarten88• 12 years ago

RavenHQ was recently added as an Azure Add-On, you can use it now. But RavenHQ seems much more expensive than DocumentDB (even when doubling the "introduction pricing"), and it's not really integrated in Azure (I understand that it runs on Amazon, in a US datacenter only)

But it will be very interesting to compare features, Raven really has a lot of advanced features.

allegory5• 12 years ago

Because that's just as much of a ghetto, Microsoft can't control it as they'd have to buy it and I doubt Oren would sell it to them.

Not only that, there are better products that are free.

pistle• 12 years ago

I'm going to posit licensing/cost issues as the easy downfall. However, something about the closed sourcing against the context of the open source movement from MS has me wondering about specific implementation details they feel are differentiating or expositions of brilliance/hacks?

allegory5• 12 years ago

No local installation. No banana.

I wouldn't tie a product to a single cloud vendor.

kolev• 12 years ago

So you're questioning AWS's success then?

cheepin• 12 years ago

nor one as expensive as azure

threeseed• 12 years ago

Absolutely.

However there is an overwhelming trend towards hosting key parts of your infrastructure including data storage. Managing a database is surprisingly difficult in particular at scale.

mariusmg• 12 years ago

Yes, because all products today are designed and for multiple cloud vendors. Come on.......

weego• 12 years ago

As the only source of DocumentDB is as an instance within Azure you are limited to that vendor. You can obviously take redis/monogo/couch and put them on any 'cloud' providers infrastructure as needs be.

Not sure if your interpretation of his point was wilfully incorrect for some reason, but it was quite obvious what he meant.

MichaelGG• 12 years ago

I wonder what the search story is. One technology that really does deliver, and has totally impressed me, is Lucene/ElasticSearch. I'm used to all sorts of hyperbolic claims, but holy shit, ElasticSearch just delivers. We tossed in about 40M documents from a SQL Server DB, and not only did it require less resources (a 30%? reduction in size), the queries are beyond anything that'd be approachable using SQL Server. And I've only touched the surface, using it as a pure plug-n-play setup.

With DocumentDB, not having a local version severely limits what I'd consider this for. Losing that flexibility is a big deal. Maybe this is just a limited preview and they haven't build the management side for local installs.

curiousDog2• 12 years ago

They released a full-text search service as well today: http://azure.microsoft.com/en-us/services/search/

nemothekid2• 12 years ago

I doubt MS will ever release a non-hosted version. I see this product as a competitive offering to Amazon's Dynamo and Google's BigQuery on AppEngine.

bigtones2• 12 years ago

I think you mean a local version, or a non-hosted version.

spicyj• 12 years ago

BigQuery isn't really in the same space; it's meant for analytics queries where speed isn't as important (queries take seconds or minutes). Google Cloud Datastore is a better parallel.

phatboyslim• 12 years ago

Similarly the Azure Search appears to be a competitor to Amazon CloudSearch. I'm not aware of a Lucene based search offering from Google, however.

johnward• 12 years ago

I haven't used elastic search but I've used lucene/solr. It's similar to a technology our startup created which was acquired by IBM and is now part of the Watson portfolio. I think our indexing engine was faster. I know for sure we were better at getting data into the index than lucene appears to be. What are the common connector frameworks for lucene? Or do you generally just push the documents to the index?

chippy2• 12 years ago

ElasticSearch benefits from Lucenes spatial support too, which is a serious benefit.

liamca2• 12 years ago

Hi MichaelGG,

I am a Program Manager for Azure Search and as curiousDog mentioned, yes along with DocumentDB, we also announced Azure Search which is a PaaS based full text search service. We actually leverage ElasticSearch at the core of this service and as chippy says about spatial search, we do have the ability to provide a pretty solid geo-spatial search capability thanks to Elastic Search and Lucene. To nemothekids's point, is it very unlikely that we will offer this as a local (non-hosted) version because we found that although ElasticSearch is awesome, one of the common complaints many admin's have is the complexity around managing systems such as ElasticSearch/SOLR/Lucene at high scale and how difficult it is to implement more advanced search capabilities such as tuning and relevancy. Those are areas we think we can add a lot of value being a fully managed service. Longer term, we think this will also allow us to bring even more value on top of search by adding in other Microsoft technologies. For example, we could tie in Bing Maps to allow you to easily tie in reverse geo-coding right into your search. Or perhaps allow you to leverage Bing's synonym list so that you could allow people to search yet find results that are synonyms to commonly searched words (i.e., user types in shoes, but in your content it is referred to as footwear). Multi-language support is actually one of the big things we want to tackle in the short term and we believe that the NLP from Office will really help jump start us with this.

Liam

tracker1• 12 years ago

We don't need the actual infrastructure or even the same software locally.. what's wanted/desired is an API compatible system that can be used to develop against. It also doesn't need to run on a local cluster for development purposes, or smaller shops.

I know everyone is all about running huge clusters behind the scenes, but most people simply don't need that, and being able to start small, and buy bigger would be a nice option.

For example, a simple node application using leveldb(levelup) for the storage interface could be very effective as a backend for development/testing. From there, you provide an API compatible interface. Open-source that version, with the disclosure/understanding that the Azure hosted version is much more robust.

I think you'd see a lot more buy in from the open-source community... even more so if you accepted PRs to make the open version more robust.

If you are at a point where you are considering the likes of ES/SOLR/Lucene etc, you are likely ready to make the jump to self-host in the cloud, or use a SaaS provider. Where people get a bit concerned is in the lock in. I know why Azure would want to present that, but I think it's a bad idea without an open implementation that allows for self-hosting for development and on the small scale.

Right now, the company I work for is hosting in Azure. I recently switched from a couple RabbitMQ queues to Azure Queues. It works fine, as was a really simple replacement for a flow of near real-time but temporary data keys. I would be much more open to using a hosted MongoDB from MS than DocumentDB, in much the same way I'm happy to see you guys embracing Redis for a cache system.

luuio4• 12 years ago

A quick comparison between DocumentDB vs MongoDB: http://daprlabs.com/blog/blog/2014/08/22/azure-documentdb/

spiderPig888• 12 years ago

Wow, how'd you get one done so fast? Were you a private preview customer?

jeroen• 12 years ago

OP did not write that. See here: https://news.ycombinator.com/item?id=8209525

cbsmith4• 12 years ago

'cause what the world needs is another proprietary NoSQL solution.

PerfectElement• 12 years ago

As long as it offers something new, I don't see anything wrong with that.

I've been working with MongoDB for the last year and DocumentDB features look very interesting to me.

cbsmith4• 12 years ago

Yeah, once you go down that road for a while, you'll understand how proprietary NoSQL is a giant PITA.

james2vegas2• 12 years ago

'cause what the world needs is another pXrXoXpXrXiXeXtXaXrXyX NoSQL solution.

cbsmith4• 12 years ago

Are those X's supposed ^H's.

hippich• 12 years ago

I don't think it is about world. I would expect it is about http://en.wikipedia.org/wiki/Embrace,_extend_and_extinguish

And looking at their diagram[1] I already see where they can make this strategy work. While their collection/document/attachment structure is quite straightforward to move to another document DB, their UDF, sprocs and triggers I bet will be microsoft-specific. And microsoft will make everything possible to "optimize data fetching by running processing code close to data" and lure developers into using these. And in the end locking software solution to microsoft platform.

[1] http://azure.microsoft.com/en-us/documentation/articles/docu...

petilon• 12 years ago

So does it run on a cluster? If so which of Consistency, Availability and Partition tolerance does it NOT offer? (See CAP theorem)?

falsestprophet• 12 years ago

It supports tunable consistency

http://azure.microsoft.com/en-us/documentation/articles/docu...

semaphoreP• 12 years ago

I don't think the CAP theorem is so black and white: http://en.wikipedia.org/wiki/CAP_theorem#2012

curiousDog2• 12 years ago

They can all be provided to varying degrees. But usually most distributed systems sacrifice some availability to guarantee consistency (i.e avoid data loss at all costs)

chippy2• 12 years ago

Spatial queries and indexing. Most data has some location component. I didn't see anything with this. Is it in there, or planned?

seanp2k2• 12 years ago

Interesting: https://github.com/Azure/azure-documentdb-python (it's empty for the moment, but glad to see first-party support for Python)

luuio4• 12 years ago

The .NET & JS repos are empty, too. I think they just have not published the code. It's still in Preview after all.

https://github.com/Azure/azure-documentdb-net

lubos2• 12 years ago

What are the limits of DocumentDB? You know, like max size of database, max size of document, max number of documents per database, max. number of attributes per document, max. number of databases per DocumentDB account.

What's the max. duration of database query, max size of query result.

What kind of performance can be expected, does it decrease as the size of database increases or it remains constant?

I'm going to wait a few days until hype settles.

reubenbond7• 12 years ago

Great suggestion, please vote here to make it a reality: http://feedback.azure.com/forums/263030-documentdb/suggestio...

majorsc2noob• 12 years ago

Most of this is documented since start. Did you actually read the docs?

http://azure.microsoft.com/en-us/documentation/articles/docu...

lubos2• 12 years ago

I did read documentation but couldn't find it. I'm still not sure how did you find it other than going into their GitHub repository.

Great find though. Thanks.

Nux• 12 years ago

A new "cool", locked-in service served on a silver platter by Microsoft to the brainwashed.

Everybody else uses open source on premises or their cloud of choice.

poolpool• 12 years ago

I wonder if this is built on JetDB

james2vegas2• 12 years ago

On Jet Blue (ESE)? It wouldn't be the first, RavenDB http://ravendb.net is based on ESE.

I remember awhle back doing work on a system that (ab)used MS Exchange Server 5 as a database, mostly because of the Outlook integration.

gamesbrainiac• 12 years ago

I find it surprising that DocumentDB wasn't already a copyrighted name. ;)

reubenbond7• 12 years ago

The @DocumentDB twitter links to a tutorial on DocDB: http://www.documentdb.com/sql/tutorial

fineline2• 12 years ago

"All JavaScript logic is executed within an ambient ACID transaction with snapshot isolation. During the course of its execution, if the JavaScript throws an exception, then the entire transaction is aborted."

Have I missed something, or have MS delivered a novel and valuable feature? I'm not aware of support for transactions across documents in other NoSQL platforms. I'd be grateful if someone has any experience or better information in that regard, thanks.

int642• 12 years ago

You are missing AmisaDB. Supports Multistatement transactions. http://www.amisalabs.com/

fineline2• 12 years ago

No I didn't know about that, thanks. I guess I was more thinking about run-on-your-own-hardware/cloud - which I know is not equivalent to the MS offering. I see orands post about FoundationDB, which also looks interesting. Seems like it would be a killer feature for a Mongo/Rethink/Couch type system.

tristanz• 12 years ago

It appears that AmisaDB is in-memory only, only supports read committed isolation, and has limited transactions semantics (no arbitrary JS). This is pretty far from what DocumentDB seems to offer.

http://www.amisalabs.com/AmisaDB_Docs.html

nikon• 12 years ago

RavenDB http://ravendb.net/docs/client-api/advanced/transaction-supp...

savecastel• 12 years ago

FoundationDB https://foundationdb.com/

Quite an amazing piece of technology, check how they test cluster failures, there is a blog post about it :)

orand• 12 years ago

If I understand correctly, their multi-document ACID transaction support is a big deal. The only other NoSQL/NewSQL systems I'm aware of with that ability are FoundationDB and Google Spanner/F1.

int642• 12 years ago

So does AmisaDB. http://www.amisalabs.com/

platform• 12 years ago

yes, it is a big deal MarkLogic does this too (I think they have been doing it since '03 or something like that)(

http://www.marklogic.com/blog/acid-transactions-check/

hackerkushal• 12 years ago

THIS THING IS A BEAST!! It is absolutely bad ass

whalesalad• 12 years ago

I liked everything about it until I saw the API for the Python client. What a catastrophe.

I pray Microsoft is looking for Python developers: https://gist.github.com/whalesalad/2142f0075c6896f4547c

ayrx• 12 years ago

Wow, the function naming alone is terrible.

    for i in range(len(path_parts) - 1, -1, -1):
        if not path_parts[i] in resource_types and path_parts[i] in resource_tokens:
            return resource_tokens[path_parts[i]]

This bit of iteration is painful to look at as well.

jkldotio• 12 years ago

Some optimisations to get it to web-scale on TempleOS were obviously necessary.

maaaats• 12 years ago

I didn't compare the code, but maybe it's autogenerated from a more C-like language, making a bit awkward syntax for a language like Python?

blablabla123• 12 years ago

The Client APIs are kind of free: https://github.com/Azure/azure-sdk-for-python/blob/master/LI...

mallipeddi• 12 years ago

What are the size limits on a collection? Docs mention transaction support is offered only within a collection. Is a collection essentially limited to a single physical machine in the background or does it span across machines? It looks like in Standard Preview, the max collection size is 10GB.

nandkishiee• 12 years ago

Sick! Love it

sarciszewski• 12 years ago

Leave it to Microsoft to give it the most generic sounding name possible.

josteink• 12 years ago

Yeah. Instead of giving it a reasonable, obvious name they should have gone with something hipstery and vague instead, like UberDcmntor.io.

No thanks. I think this will do just fine.

rtkwe• 12 years ago

No no no, it also has to be just one common word so that it's completely ungoogleable by itself.

DocumentDB

Related Stories