ClearDB

In the blog post announcing the ClearDB add-on for Heroku it states:

ClearDB is an affordable and powerful database-as-a-service solution that provides geographically distributed, highly available MySQL database services to the cloud.  With ClearDB, you will stay connected to your database even if entire regions of the Internet experience sudden, pervasive service outages.

Cool, I get it, ClearDB is making it database replication and high availability easy. But what’s missing from the ClearDB web site is a statement that explains, in detail, data access jurisdiction issues that you should understand and consider before you go replicating your data all over the world.

Granted, cloud data jurisdiction is not an issue that’s specific to any one cloud data storage solution such as ClearDB — I don’t mean to pick on ClearDB. Nonetheless, I believe that all such vendors that provide you the ability to store your data in the cloud need to give us more information and a better understanding about such legal considerations so that we can make wise decisions for our businesses.

Need a full-featured, ANSI SQL compliant database? If you don’t want to wait for your IT department to provision it for you, or don’t want the hassle of installing and configuring a whole bunch of software on your local machine, then cloud database services are a great option for you. And today, you can now access a new DbaaS from Heroku called Heroku Postgres.

[Full Discloser: I work for salesforce.com, the parent company of Heroku.]

PostgreSQL

Heroku Postgres is built using PostgreSQL, commonly known as Postgres. Most of my regular readers probably already know about Postgres, but for those of you that don’t, it’s a terrific open-source database management system that can satisfy the needs of even the most demanding apps. Among the many features you get are ACID and ANSI SQL compliance, data replication, full text search, and much more. It’s also developed by a global community of developers, which means that Oracle can’t buy it I suggest you read here for a complete list of Postgres features.

Heroku Postgres

Heroku Postgres is more than just Postgres as a service. Sure, you get all of the things you’d expect from a DbaaS offering that let you focus on app development rather than keeping the lights on, features such as effortless self-service database provisioning, simple yet secure access to your database from anywhere, automated backups, redundant transaction logs, and expert service health monitoring. And considering that Heroku has been operating Postgres for a good long while as part of their original platform as a service, they’ve got the experience to make Postgres reliable. But wait, there’s more.

The folks at Heroku are innovating with Heroku Postgres. How so? Consider Heroku’s Fork and Follow features (currently in Beta).

• Forking a database is like forking source code — it’s a simple way to create a clone of your source database. You can then use a database fork to test new schema migrations, perform load tests, or any operation that you want to isolate from your production database. When you are done with your work, simply trash your database fork.
• Read-only database replication is a common approach that many apps leverage to offload CPU intensive query processing from the primary, transaction processing database to one or more read-only replica databases. However, database replication is notoriously complex to configure and maintain. Not so with the Heroku Postgres follow feature. With a single command, you can create a read-only database follower (replica), which the service then keeps in sync with your primary database. Your apps can then use database followers to process queries with up-to-date data.

Getting Started

To get started with Heroku Postgres, head over to the main Heroku Postgres web site. Sign up, pick your plan, and then check out the documentation for more information about this new service.

First, the paper contains a lot of good concepts that you should know about if you are thinking about DbaaS. There’s also some good discussions about cloud database service from the point of view of several different types of actors.

Predictably, the earlier parts of the paper are setup material for marketing Oracle’s products, namely Oracle Database and related technologies like Oracle Enterprise Manager and Oracle Exadata Database Machine. I don’t have any problems with this, but just beware of what you will be reading.

The paper doesn’t contain any material about the newly announced Oracle NoSQL strategy. I’m left wondering how that technology will integrate with the architectures that Oracle is recommending.

What’s missing is meaty, how-to type material. I understand this is a white paper, so the audience is more the decision maker than the implementer. But it would have been nice to have references to other documents or online resources that could have walked me through the implementation of their suggested architectures.

[Update 10/11/2011: I just found the Chapter 8 in the Oracle Enterprise Manager Cloud Administration Guide.]

I hope that you find this review helpful.

[Cross posted at The Cloud View]

I’ll plan on doing a post on each product when I get some time. But for now, here are some quick summaries to help get you caught up.

Database.com

Database.com is generally available [full disclosure: I work for salesforce.com]. Announced last December, Database.com is now officially open for business. You can sign up and get started for free! Make sure to see the DevCenter for a long list of resources that will help get you started. Personally, I think the don’t miss piece to read is the paper Architecting Database.com Apps, A Design Primer because it gives you a feel for the technology behind the marketing messages, including open, social, mobile.

VMware vFabric Data Director

From VMware’s Web site, “vFabric Data Director is an enterprise solution that extends the benefits of VMware’s leading virtualization platform to the database tier and lets administrators securely automate and delegate routine tasks, including database provisioning, backup, and cloning. The first database supported on Data Director is VMware vFabric Postgres 9.0 (vPostgres), a new offering from VMware based on and fully compatible with PostgreSQL.” There’s a nice data sheet that gives you some great background on the product.

A couple of things I focused on while reading the data sheet include:

• The datasheet states “Robust Role-Based Access Control (RBAC) restricts system access to authorized users. Granular privileges enable IT administrators to control which users can take what actions at a fine-grain level.” Sounds great, but the statement lacks detail. A quick scan of the Postgres 9.0.4 documentation tells me that you can do some field-level security. But if you want row-level security, I think you’ll have to build it yourself.
• Elastic Database Memory seems like a nice feature. “Patent-pending technology integrates the database memory structures with vSphere to finally enable databases to fully leverage vSphere’s virtual resource capabilities, while maintaining consistent database performance even with unexpected workload spikes.” Not sure why the word “finally” is included in that description?
• Also nice, “Smart Database Configuration – vFabric Postgres 9.0’s smart database configuration algorithm automatically derives optimized database configurations without involving DBAs.” I’d like to test that one out personally, as I’ve run across many a DBA in my day that couldn’t configure systems to save their lives
• “Scalable, elastic, multi-tenant”. I see these terms thrown in marketing pieces all time without much explanation. Drives me nuts! I need some details here. How does it scale? Does it scale for both read and write throughput? Multi-tenant? How so?

EnterpriseDB Postgres Plus Cloud Server

From the EnterpriseDb Web site, “Postgres Plus Cloud Server is cloud provisioning software that automatically scales out PostgreSQL and Postgres Plus Advanced Server databases (i.e. upgrade CPU class, add storage, and manage replicas) to provide DaaS (Database as a Service) in both public and private cloud environments.” Other than getting the wrong acronym for Database as a Service (it’s DbaaS, not DaaS, which stands for Data as a Service), the product description is exciting. The product is in Private Beta at the time of this writing, and currently supports both Amazon and Eucalyptus private clouds.

The folks at EnterpriseDB have been doing Postgres for some time now, so I’m sure they’ll do a fine job of taking Postgres to the cloud. There is some great background information at the site, so check it out. Best I can tell is that they are supporting Postgres v8.3. And the replication for scalability is read-only replication — that’s fine for some apps, but not for high-volume OLTP apps.

Heroku Postgres

Ok, another Postgres cloud database service. Seeing a trend here?

Heroku [full disclosure: Heroku is part of salesforce.com, the company that I work for] has also been doing the Postgres thing for quite a while now, so they’ve got it down pretty well too. From the Web site, “A powerful, reliable, and durable open-source SQL-compliant database, PostgreSQL is the datastore of choice for serious applications. Now it is available in seconds with a single click . Never worry about servers. Never worry about config files. Never worry about patches. Simply focus on your data.” A couple of innovations to check out are the Fork and Follow features, something I haven’t seen elsewhere.

Xeround Add-On for Heroku

During its Beta, I tried Xeround and really liked it. These folks have done a nice job of making it easy to provision and use MySQL in the cloud. For example, tied into the Xeround UI you have access to the same database administration/data browser tool most of us are already familiar with — phpMyAdmin.

Now Xeround is expanding its reach to different platforms. In addition to availability on Amazon AWS and Rackspace, you can now tie Heroku apps to Xeround databases. Pretty cool if you like Heroku.

NuoDB (formerly NimbusDB)

NimbusDB, covered several times before in this blog, is now NuoDB, and has moved into Beta 2 stage. Personally, I like the old name better. By the way, is that old URL for sale Jim?

Scalebase

ScaleBase 1.0 is now available. Scalebase helps you with MySQL sharding by implementing a SQL load balancer in front of your servers. Although not a DbaaS itself, I include this product in the discussion because Scalebase is available as a service (as well as a downloadable product).

Next up: A full DbaaS Product Directory update with the information related to all of this great news!

At MongoHQ, we are passionate about the future of the data layer and have a deep desire to build and provide developers with tools and solutions that not only make them more productive, but make them happier and successful.

I’ll have to update the DbaaS  product directory this weekend.

Steve

• On Monday, July 4, at the “Cloud 2011 Summer School”, I’ll be presenting the “Cloud Development and Deployment” tutorial. Designed for developers by developers, this workshop mixes presentations with hands-on coding exercises that teach you how to build a cloud app using the world’s premier cloud platform, Force.com. This tutorial will provide you with a deeper understand of cloud computing and help you build a fully functional cloud app that you can extend on your own. Three full hours of hands-on training, workbook included!
• On Wednesday, July 6, I’m presenting my paper, “Optimal Multitenant Designs for Cloud Apps.” Often overlooked by application architects, multitenancy is the core technology of a cloud that controls how it shares computing resources among applications. Two basic approaches for multitenancy will be discussed, virtual and organic multitenancy, along with the benefits and disadvantages of each approach. Then you will discover why the database that supports a cloud app is a critical consideration, and learn about emerging cloud database options that promise to simplify the development and management of apps executing in the cloud. Along the way, we’ll draw conclusions about which types of solutions best fit particular application scenarios.

Hope to see you there. Please tweet me @sbob909 if you have any questions.

Here’s the official billing:

One of the biggest challenges that remain is managing data in the cloud. While the need for database scalability, reliability and performance has always been an issue in computing, the cloud introduces new challenges such as elasticity, geographic distribution and multi-tenancy. It has become a breeding ground for a wave of new solutions, including database-as-a-service offerings. In this session, we will engage in a discussion with some of these cutting-edge companies. We’ll define what is a “true” cloud database is and explore the unique challenges of managing data in a cloud environment.

Moderated by:

Geva Perry - Cloud Strategy and Marketing Advisor, Blogger, Thinking Out Cloud

Speakers:

• Mike Miller - Founder and Chief Scientist, Cloudant
• Razi Sharir - CEO, Xeround
• Jim Starkey - CTO, NimbusDB
• Barry Zane - CTO, ParAccel

My summary:

• Yawn. No fireworks, at all. Come on guys, mix it up a bit, make things memorable!
• Geva Perry was well-prepared and did a nice job as the moderator.
• Mike Miller came off as the smartest guy on the panel, in my opinion.

Judge for yourself. Please let me know if you disagree with my opinion.

[Cross-posted at The Cloud View]

Join me and Jim Rivera on June 3, 2011 for a sneak peek at Database.com. You’ll leave with a great understanding of Database.com and see some exciting demos of its key features. Don’t miss it! Register today.

There are no chapters/bookmarks in the video, so here’s a talk timeline to help you find what you are looking for …
0:0-0:31: All Clouds are Not Equal

• IaaS & PaaS, defined and compared
• multitenancy defined and why it’s important
• how multitenant architectures fuel rapid app advancement and innovation
• Q & A

0:31-1:27: Experiences Building Database Multitenancy

• competing challenges you’ll face
• the multitenancy spectrum: architecture options for tenant data isolation
• some relational and non-relational table design options
• system-level architecture challenges
• Q & A

1:27-End: The Multitenant Architecture of Database.com

• overview
• Database.com’s multitenant data model
• Database.com’s multitenant processing architecture
• proof of the power of multitenancy
• Q & A