In-Depth

Rx for Enterprise Information Integration Woes

Avaki's Data Grid 4.0 is designed for highly distributed environments

• By Stephen Swoyer
• 09/24/2003

This July, the enterprise information integration (EII) market got an infusion of fresh blood when Avaki Corp., a grid-computing pioneer, introduced its Data Grid 4.0 product. A data aggregation platform designed for highly distributed environments, Avaki maintains that Data Grid 4.0 is just what the doctor ordered for ailing enterprise EII efforts.

We spoke recently with Craig Muzilla, vice president of marketing and strategy with Avaki, about why his company’s grid computing roots give it a competitive leg up over other vendors in this space. The upshot, Muzilla maintains, is that grid computing vendors have already solved many of the access, provisioning, and security problems that EII vendors are only now starting to address.

Why do you believe that grid-computing technology is a good solution for enterprise information integration efforts?

If you look upon it, [grid computing is] coming up with a global name capability of identifying resources and helping you to use those resources. This also could be used as a virtualization layer for identifying data resources and software resources [and facilitating access to them]. The common knowledge out there in the marketplace is always associated with compute grids—even the big guys [IBM, HP and Oracle]—but people are beginning to identify other uses for this virtualization capability. What we found was that people were struggling with how they identified the data resources and got the data to the grid.

And you believe that these problems parallel those that enterprises typically encounter in EII initiatives?

They’re basically the same problems. Things like: how do people get to the data without having to replicate it, or without having to set up extra storage devices? How do they get to the data when there are all these ownership and authentication issues for different data resources? So we began hearing that over and over and made a decision that we thought that this data issue was a more interesting issue, more strategic to the company.

You’re shipping a product [Data Grid version 4.0] that you’ve touted as an ideal tool for EII. Is this your first data integration offering, or is this something that you have prior experience with?

We added data grid capabilities into the compute grid offering [that we previously shipped], and [Data Grid] 4.0 just shipped in July. But data integration has always been a part of the product. We came out with a J2EE-based product in the fall of last year, and that was focusing on data integration problems, [such as] how do you provision unstructured data or flat file data across an organization. Now, with 4.0, we’ve added relational capabilities, so that you can set up a SQL statement or a stored procedure and bring that data into the grid, cache that, do manipulation or aggregation of the data. Things like that.

You’re a data integration outsider, however, in that you’re coming from the grid computing space. Why should customers take a chance on Avaki when, say, IBM will sell them the EII tools that they need?

What do we bring to EII? Well, our architecture itself is a peer-level distributed architecture, so if you look at EII goals or strategies out there, it’s a good fit, because usually [with most EII efforts], you’re using an EII layer to connect to all kinds of distributed data. One of the biggest differences in a distributed grid architecture is that the whole thing can organically grow. There’s administration, yes, but in a data grid it’s easier to break up your administration, break up your security. Using a grid, you can give local data owners the chance to manage their resources without going to a central administrator to manage security and provisioning. So the issues of federated security, the issues of performance management, caching, all of those things that you have to deal with for EII, we’ve already dealt with everything in our product.

You mentioned IBM, and the truth is that we have a pretty good relationship with IBM. On our strategic partnership front, we’re actually engaged with a lot of the BI guys right now. With IBM, they’re focusing on doing the distributed query, they’re focused on a real-time distributed query; we’re focused on provisioning the data to that query. They don’t really work very well with provisioning on a wide basis, they don’t work very well with unstructured data. We excel in both of those things. When they had their big grid computing announcements earlier this year (http://info.101com.com/default.asp?id=154), we were the only data-focused company that they selected [to partner with IBM].

Is Data Grid just a tool for EII or does it have any broader BI applications?

We’ve had a phenomenal response from people who were considering data marts. I mean all of these professionals who are considering data marts but are turned off by some of the problems associated with them. It seems that there is a pain point that is still pretty acute out there with trying to provision data, so an easier way to get to the data and provision it seems to resonate with a lot of people who are considering [a data mart].

More generally, companies mostly at the CIO level are looking at their organizations, and there are many initiatives out there to put some kind of layer in an organization where you can get control of the provisioning capabilities, because not only does it make it easier to publish the data out of a database, it makes it easier to track it.