Azure Goes Database Crazy With One New NoSQL, Two New SQL Services

An anonymous reader quotes a report from Ars Technica: In its continued efforts to make Azure a platform that appeals to the widest range of developers possible, Microsoft announced a range of new features at Build, its annual developer conference. Many of the features shown today had a data theme to them. The most novel feature was the release of Cosmos DB, a replacement for, or upgrade to, Microsoft's Document DB NoSQL database. Cosmos DB is designed for "planet-scale" applications, giving developers fine control over the replication policies and reliability. Replicated, distributed systems offer trade-offs between latency and consistency; systems with strong consistency wait until data is fully replicated before a write is deemed to be complete, which offers consistency at the expense of latency. Systems with eventual consistency mark operations as complete before data is fully replicated, promising only that the full replication will occur eventually. This improves latency but risks delivering stale data to applications. Document DB offered four different options for the replication behavior; Cosmos DB ups that to five. The database scales to span multiple regions, with Microsoft offering service level agreements (SLAs) for uptime, performance, latency, and consistency. There are financial penalties if Microsoft misses the SLA requirements. Many applications still call for traditional relational databases. For those, Microsoft is adding both a MySQL and a PostgreSQL service; these provide the familiar open source databases in a platform-as-a-service style, removing the administrative overhead that comes of using them and making it easier to move workloads using them into Azure. The company is also offering a preview of a database-migration service that takes data from on-premises SQL Server and Oracle databases and migrates it to Azure SQL Database. Azure SQL Database has a new feature in preview called "Managed Instances" that offers greater compatibility between on-premises SQL Server and the cloud variant, again to make workload migration easier.

Dear microsoft

[Anonymous Coward]

The Check cleared the bank, thanks!

Yours, Slashvertisement team

Re:

[art123]

Or it could be that the annual Microsoft developer conference started today so there are a lot of announcements coming out, some of which are making their way to the Slashdot front page.

Re:

[nyquil superstar]

I mean, I get what you are trying to do. But doesn't the fact that Microsoft is hosting their big, annual, worldwide developers' conference and announcing things that would be of interest to tech folks qualify as legitimate news?

Cosmos DB new database Azure only?

[JcMorin]

If I want to try out Cosmos DB I have to pay an Azure subscription? As I can see, it's not a stand-alone product at all.

Re:

[Tablizer]

They don't have "Cosmos Express", like their crippled "Express" version of MS-SQL-Server? (One core, limited size, no easy way to back-up, & missing misc. features.)

Re:

[Tablizer]

To back up Express, the script has to shut it down (stop the process), copy the data files, and then restart it. That's risky.

Re:Cosmos DB new database Azure only?

[phantomfive]

Amazon has a similar database in AWS (DynamoDB). These kinds of products are not a good idea unless you want to lock yourself in to the platform (or unless you know how to write a really, really good abstraction layer).

Re:

[vipw]

On the other hand, the products have such a simple interface that fairly trivial to make an abstraction layer.

Re:

[luis_a_espinal]

On the other hand, the products have such a simple interface that fairly trivial to make an abstraction layer.

I was going to say that. Good abstraction layers are not that hard if the system in question is layered and compartmentalized accordingly and data is modeled appropriately. And with things like Spring Data (at least in Java land), it's almost trivial. It only gets icky in edge cases where things like throughput are paramount (or the data models are wrong for the type of abstraction in use, like trying to use a document database for time series data.)

Re:

[javaguy]

Back when I was a junior developer I used to write abstraction layers in case we wanted to change databases. In 20 years no system I worked on ever actually changed. However, in modern times, moving between cloud providers may make this more likely, but still probably relatively unlikely.

If you change from (say) Azure to AWS it's going to be a significant effort. Having to change an application would probably be a small component of this, given the code behind the abstraction layer would have to be changed

Re:

[spongman]

apparently MS's CosmoDB allows you to talk to it using the MongoDB APi (among others). so, in theory at least, you can transition to/from it easier than if it were using a custom API like DynamoDB. i haven't played with it myself yet, though. this is just based on watching the keynote and a brief perusal of the high-level marketing docs.

Re:

[phantomfive]

I can't speak for this one, but in my experience Microsoft's Entity Framework works really well with MSSQL, but not so well with MySQL. YMMV.

Re:

[TheSunborn]

Oh my god I hope not. We host our servers at aws, and I really hope they provide better reliability then Microsoft.

Re: Good

[buchanmilne]

Really?

Browsing
https://azure.microsoft.com/en... [microsoft.com] makes it look like running 3 Openstack deployments would be easier (and of course cheaper) than keeping something running on Azure ...

Planet- scale apps

[Anonymous Coward]

Planet-scale, eh? I can tell you already the answer is 42.

Re:

[deKernel]

Well of course it is "planet-scale" because they were able to fully future-proof the design.

In a way, I feel sorry for the Azure team...

[QuietLagoon]

They are stigmatized by the egregiously bad actions of the Windows team, the malware-like tactics that were used to force Windows users to upgrade to Windows 10 against their apparent will or knowledge.

.
Azure may be the greatest thing since sliced bread, but it will always suffer from the anchor that is the Windows 10 update disaster.

Re:

[hackel]

You shouldn't feel sorry for them. They all *chose* to work for such a vile company. It's not as if Microsoft's history is a secret. They knew perfectly well what they were getting into. And honestly, Windows 10 was the *least* of Microsoft's problems over the years. People are whining and complaining over nothing.

Really? MS hosting Planet-Scale databases?

[grep -v '.*' *]

Planet-Scale databases? Sounds like they've already started loading their database with everyone's Windows 10 data and metadata. Since they couldn't be bothered to generate any FAKE data, this one's considered a "live fire" test.

Don't you just hate it when the Whole Planet BSODs though? It takes real expert to generate that sized mess.

?? So can you generate a BLOB the size of a planet? If so, that would explain the old joke about "You can't own everything in the world since there'd be no place to put it." I guess tape backups really ARE dead.

Re:

[zifn4b]

Planet-Scale databases? Sounds like they've already started loading their database with everyone's Windows 10 data and metadata.

Nope... they've started loading it with my distributed peer-to-peer statistical noise data generation system. Muhahahaha. Big data. Yep, it's big, a big steaming pile of useless... bits.

"Database crazy"?

[rebelwarlock]

I'm not going to read articles written by children.

As a DocumentDB developer

[Anonymous Coward]

Don't use this piece of garbage.

Performance is abysmal, cost sky-high (for decent read/write performance in a single collection, you're looking at over $1,000/mo, and that's for data sets under 10 GB because if you go over 10 GB it forces you to partition the data, which dilutes performance between the partition shards), there is no proper tooling (unless you count the Azure "portal" which is a series of perpetual slow or failed AJAX requests), poor SDK, and lacking basic features (e.g. update part of a doc

The best part about standards...

[xxxJonBoyxxx]

>> Azure Goes Database Crazy With One New NoSQL, Two New SQL Services

The best part about standards is that there are so many different ones to choose from. Wheee!

Meanwhile in the real world

[Khyber]

I just use KirbyCMS, which is faster than any database because it runs off the native filesystem, and provides me tools that most databases simply cannot or do not provide.

Re:

[hackel]

You clearly don't know what you're talking about. A database is vastly different from a CMS. There's nothing fast about "the native filesystem." It might work fine for running your blog, but for any serious database application, it's a non-starter. You honestly can't even compare the two.

Re:

[Khyber]

" A database is vastly different from a CMS. There's nothing fast about "the native filesystem." It might work fine for running your blog, but for any serious database application, it's a non-starter. You honestly can't even compare the two."

No, really, it's not, if you've bothered using every single one out there. And there's plenty fast about the native filesystem. You must not know much about moving tiny bits of data around in a rapid manner. I've been running multiple websites (including my businesses)

Microsoft Loves Open Source!

[EndlessNameless]

Now that they found a way to charge for it, they are totally jumping on the bandwagon.

Sarcasm aside, this could be good if they contribute code for better performance or scalability to the projects. Then again, their code could be so tightly coupled to the Azure infrastructure that it is useless to anyone else. I don't really follow those projects to know if that's happening though.

Does anyone even use this?

[hackel]

Who in their right mind would trust business-critical data to Microsoft? Honestly? It makes no sense. Surely no respectable devops person would allow this. Is this 'Cosmos DB" even open source? How does anyone know what it's doing with their data? Microsoft seriously needs to get out of this market.

Re:

[spongman]

> How does anyone know what it's doing with their data

that seems like just the kind of question that these people [microsoft.com] would ask.