Part 1 of this blog series talked about my attendance of Dreamforce 2010, salesforce.com’s annual user conference, which has over the past several years become a highly anticipated and entertaining end-of-the-year fixture for enterprise applications market observers. My post concluded that while Dreamforce 2009 was mostly about continued growth of the cloud computing trailblazer and unveiling of Salesforce Chatter, the company’s nascent social platform and collaboration cloud (as duly covered by my blog series), the overall Dreamforce 2010 theme was cloud proliferation (and salesforce.com’s further diversification).
In his blog post, Louis Columbus states that at the center of Dreamforce 2010 was the transformation of salesforce.com into an enterprise cloud platform provider, starting with endorsing open application programming interfaces (APIs) including REST (Representational State Transfer), which its developer community had reportedly been requesting for over a year. Moreover, after realizing the proprietary nature of its Force.com cloud platform (and its Apex code), salesforce.com CEO Marc Benioff and his co-founder Parker Harris have recently decided to decouple Force.com into a more open application layer, for platform as a service (PaaS) purposes and a database layer for providing infrastructure as a service (IaaS).
Database in the Cloud
The latter layer was named database.com, which should help to attract Java, Ruby on Rails, PHP, Perl, etc. developers, who have likely perceived Force.com (Apex) as non-mainstream technologies. When it becomes generally available (GA) in early 2011, customers will be able to use database.com as the back-end for both on-premises and cloud-based applications. Open APIs will also provide compatibility for applications developed for smartphone and mobile devices.
We all learned at the Dreamforce 2010 media lunch that the “database.com” name was Benioff’s failed entrepreneurial attempt between his tenure at Oracle and his salesforce.com run (some time in 1999). Indeed, from the light of salesforce.com’s resounding success, it is difficult to imagine Benioff failing, but it apparently happens to the best of us. At least, the dormant domain name will come in handy now, since database.com or “Database as a service (DaaS)” might be just what the doctor ordered for mobile and social applications, given that traditional on-premise databases are not easily accessible since they are usually behind firewalls.
Cloud database is not salesforce.com’s invention per se, given the existence of Amazon’s SimpleDB, FathomDB, LongJump, Xeround, FlatDb by Google, and Microsoft’s ongoing work on SQL Azure. But most of these offerings are only good for creating a few very large database tables (so-called “fat tables”) to be stored in the cloud.
Conversely, database.com is an enterprise-class relational database technology for a vast number of tables, relations, database normalizations, etc. It is largely built on top of Oracle’s relational database technology, although salesforce.com has added a number of other traditional on-premises databases into its datacenter portfolio: the Dreamforce 2010 demo featured a recruiting application on top of a Progress OpenEdge database. The cloud database offering can be hosted on multiple IaaS offerings, including Amazon’s Elastic Cloud 2 (EC2), Microsoft Windows Azure, and salesforce.com’s own infrastructure.
Benioff asserted that the scalability offered by the cloud, together with database.com’s affordable pricing, will make database.com an attractive proposition for its customers. Some observers have suggested that the offering would put salesforce.com in competition with Oracle, IBM, Progress, and other on-premises database providers.
In truth, it rather makes the vendor a reseller of these on-premise databases. As a matter of fact, during the media lunch at Dreamforce 2010, Steve Fisher, the chief architect at salesforce.com, encouraged Oracle and IBM to offer their databases in the cloud, and even had a few good words for what Microsoft has been trying to do with SQL Azure (although one should keep in mind that SQL Azure and its on-premises SQL Server counterpart will not seemingly be in the same code).
The Chatter Buzz
The media and analyst session discussed, inter alia, Salesforce Chatter’s state of affairs. The social platform will seemingly benefit from the GroupSwim recommendation engine, which was tacitly acquired in late 2009. The engine will, e.g., provide recommended items, most popular items, etc. (similar to the Facebook or Twitter recommendations to which we are all accustomed to by now).
File sharing (and the consequent email attachments reduction) is very important for Chatter adoption and the idea is for the platform to be more functional and easier to use than the Dropbox service. Some of the upcoming features will be as follows: mentions, like, file sharing (user to user), topics widgets on profiles, trending topics, comment by email, feed search, etc. Again, many of these features are recognizable from our own use of Facebook and Twitter.
Chatter and its mobile version have already been available free to all existing salesforce.com customers, as opposed to previously being available only to customer relationship management (CRM) users. These CRM users can now send viral invitations to join Chatter discussions to all of their colleagues within the company (which is hoped to increase the use from typically 20 percent of users to everyone using it). Chatter.com is slated for a February 2011 release as a free public version (Facebook for enterprises of a sort), but the details on its capabilities and envisioned purposes are still sketchy.
Last but not least, there is a standalone social media platform version, Chatter Plus, for US$15 per user per month. But in my opinion, Chatter still needs the presence information, instant messaging (IM), blogs, wikis, polls, and a number of other capabilities to be a full-fledged standalone social business platform in the league of Jive Software, Saba Live, Socialtext, and even Microsoft SharePoint 2010 and IBM Lotus Connections.
In addition, Chatter is still intended for intra-enterprise (within a firewall) collaboration only. The recent post-Dreamforce acquisition of the videoconferencing vendor DimDim should mitigate many of the above issues.
Taking Stock of salesforce.com’s Clouds
At the Day One of Dreamforce 2010, Benioff outlined the following six salesforce.com cloud offerings:
Benioff hinted that the Day Two would unveil two more cloud offerings, and that Force.com would be broken down into several cloud applications building flavors. To that end, there are the following application development environments that cater to different user constituencies and developers’ language preferences:
Is Ruby That Precious a Gem?
The latter PaaS piece came from the very recent acquisition of Heroku for US$212 million, a San Francisco start-up of 30 employees (that apparently are worth US$8 million each to salesforce.com) that provides a hosted development environment for applications built in the Ruby on Rails programming language. Ruby is touted as a true cloud computing technology with the following features: rapid development, productive programming, amenability to mobile and social applications, and massive scalability.
At present, Heroku hosts more than 106,000 applications on its platform, which is based on Amazon’s EC2 cloud environment. The company features Best Buy and FlightCaster as its customers and success stories. All of this, salesforce.com hopes, will help unlock the Force.com offspring offerings’ potential as an all-purpose collective cloud platform for business.
Finally, the latest (the eighth for now) of salesforce.com’s cloud applications is RemedyForce, announced in alliance with BMC Software for IT service management (ITSM) capabilities. Dear readers, what are your comments and opinions with regards to salesforce.com’s diversified PaaS strategy and its expansion well beyond the realm of CRM?
Nice article! However, I have to correct you on one item. LongJump is not a database table architecture. We use a completely relational model, supported by metadata which infuses SQL tables with common enterprise/business frameworking so the data has “value-add” functionality (e.g. security, processing, formatting, etc.).
You are right, however, in saying that database.com is not new. Salesforce.com is simply exposing their existing technology into a lower level that is DBA-friendly.
Thanks for your input, Derek.
I did say “most of these DB offerings” and thanks for clarifying about LongJump.
Excellent article, P.J., very thorough and thanks for your freely sharing your knowledge and insights. Thank you for mentioning my post on the transformation of Salesforce.com as well. I look forward to your future posts, and have a great weekend.
took a look at the demo and it does look impressive. We also don’t know how this will afcfet Seesmic’sa0announcement of their CRM for Salesforce Beta app but it may quickly reduce the need of Seesmic’s app when users can use touch.salesforce.com