By Michel Prompt, Founder & CEO
Last week’s webinar on “Data Clouds”, featuring Lyn Robison, Research Director for Data Management Strategies at Burton, is already generating a buzz in the blogosphere. In his recent blog post responding to the webinar, Jnan Dash expertly captured the data cloud concept. Formerly a senior executive at IBM and Oracle, Jnan’s insight into the architecture and potential of data clouds is impressive.
You’ll Need More than SQL in the Cloud
As Jnan mentioned, our background at Radiant is in directories and LDAP/X500. And in a world dominated by SQL and relational databases, one might wonder what a hierarchical data model could bring to the equation. Didn’t we all learn back in Database 101 that relational is the mother of all hierarchies and networks? Of course, I am being a bit disingenuous here, because unless you live on another IT planet, the buzz around an alternative model to straight-jacket SQL is growing stronger by the day. Just look into NoSQL or check out the excellent post in the identity space by my friend Dave Kearns.
Where SQL Falls Short: Delivering the Final Experience in Real Time
As we all know, SQL databases offer tremendous flexibility in joins and updates, but can require infinite time and patience to retrieve your data. And in a world where everything—including rich context—will be digital and information will be widely distributed, that’s a real problem. Sure, we could potentially generate every view we want with SQL through distributed joins. But in such a virtual world, the hero of the SQL Holodeck would never get to kiss the girl, because the next image would be frozen in an eternal distributed join. Although many companies keep trying to buy the “ultimate SQL engine,” spending millions of dollars to close the gap, this is just the wrong approach.
Think about the difference between rendering the image for a special effect and then freezing the frame in the final movie—it’s is a whole different set of tools. While SQL is great for “generating the images” offline, you still you need another architecture to deliver the final experience in real time. Hence the emergence of cache, NOSQL, and other alternative models beyond the SQL layer, and the need for another service—a data service, a data cloud—on top of the existing application infrastructure.
The Best of Both Worlds: Identity and Context Virtualization
By virtualizing and caching identities and queries, RadiantOne combines the flexibility of a SQL database with the speed of a directory. Our technology combines the best of both worlds to allow enterprises to link cloud-based and existing infrastructures, and smoothly transition into the cloud.
Jnan also questioned the deployment of the tools that will enable data clouds and context management. Considering that the possibilities border on science fiction, it makes sense to want to see results! However, early adopters such as Sony Electronics and others, as well as many agencies and US government entities—such as the Department of Defense—are already deploying Radiant Logic’s technology to link diverse data sources and provide user information in context.
A Final Point on Scaling and Cloud Deployment
Thanks to native partitioning and replication of the LDAP model, scaling is not an issue. Contrary to a database schema, a hierarchy yields to easy partitioning by design—that’s why DNS and other basic Internet infrastructure scale so well. This what I meant when I referred to “scaling at the SaaS level” during the webinar. However, because we are maniacs for performance—at the end of the day, speed is always linked to higher productivity—we are also looking to push the envelope by combining Hadoop with our LDAP smart cache engine in our next version.
So yes, Jnan is right to stress this point. But when we talk about data clouds, we are not talking only about clouds as an abstraction concept—we’re talking about clouds as a delivery layer. So the discussion encompasses not only SaaS, but also the IaaS part of the story. This picture shows how we see the roadmap for Identity and Context Virtualization in the cloud:
Finally, we’d also like to clarify that despite having a French CEO with a thick accent—who’ll become a US citizen next month!—Radiant Logic is proud to be an American company headquartered right here in beautiful Novato, California. Not that this matters much in a global economy, but we believe in truth in marketing!
Thanks again, Jnan, for your excellent post. Stay tuned for my next blog post, where I will discuss how we can a deliver a context and semantic service for the data cloud on top of your existing infrastructure of data silos.