Wednesday, June 29, 2011

Talend Open-Source Approach Provides Holistic Integration Capability Across, Data, Devices, Services

Transcript of a sponsored BriefingsDirect podcast on enterprise integration and new tools to put control in the hands of "the masses."

Listen to the podcast. Find it on iTunes/iPod and Podcast.com. Download the transcript. Sponsor: Talend.

Dana Gardner: Hi, this is Dana Gardner, Principal Analyst at Interarbor Solutions, and you're listening to BriefingsDirect.

Today, we present a sponsored podcast discussion on how the role and impact of integration has shifted, and how a more comprehensive and managed approach to integration is required, thanks to such major trends as cloud, hybrid computing, and managing massive datasets.

Moreover, the tools that support enterprise integration need to be usable by more types of workers, those that are involved with business process activities and data analysis. The so-called democratization of IT effect is also rapidly progressing into this traditionally complex and isolated world of applications and data integration. [Disclosure: Talend is a sponsor of BriefingsDirect podcasts.]

So, how do enterprises face up to the generational shift of the function of integration to new and more empowered users, so that businesses can react and exploit more applications and data resources and do so in a managed and governed fashion? This is no small task.

We're finding that modern, lightweight, and open-source platforms that leverage modular architectures are a new and proven resource for the rapid and agile integration requirements. And, the tools that support these platforms have come a long way in ease of use and applicability to more types of activities.

We're here today to discuss how these platforms have evolved, how the open-source projects are being produced and delivered into real-time and enterprise-ready, mission-critical use scenarios, and what’s now available to help make integration a core competency among more enterprise application and data activities and processes.

Please join me now in welcoming our guests today. We're here with Dan Kulp, the Vice President of Open Source Development at Talend’s Application Integration Division and also the Project Management Committee Chair of the Apache CXF Project. Welcome back to BriefingsDirect, Dan.

Dan Kulp: It’s great to be here. Thank you.

Gardner: We're also here with Pat Walsh. He is the Vice President of Marketing in the Application Integration Division at Talend. Hey, Pat.

Pat Walsh: Nice to be here as well.

Gardner: Pat, let me start with you. We're talking about a shift here in some major trends. Everyone is talking about how IT needs to react differently. There is lots of change going on. Integration has always been important, but now it’s probably more important than ever.

With some of the shifts in computing models, such as cloud and the data intensive atmosphere that most organizations are now operating in, why is integration a real issue that needs to be approached differently?

Overriding trends

Walsh: We're seeing a couple of overriding trends that have really shifted the market for integration solutions. The needs have shifted with changes in the workplace.

First and foremost, we're seeing that there is much more information that needs to be managed, much more data associated, and there are a couple of drivers of that.

One is that there are many more interactions amongst different functional units within a business. We're seeing that silos have been broken down and that there’s more interaction amongst these different functions, and thus more data being exchanged between them and more need to integrate that data.

There’s also this notion of the consumerization of IT, that with so many devices like iPhones and iPads being accessible to consumers in their everyday life. They bring those to work and they expect those tools to be adapted to their workplace. With that just comes an even larger increase in the data explosion that you had referenced earlier.

Coupled with that are overriding trends in IT to shift the burden of supporting systems away from the traditional data center and into the cloud. Cloud has been a big movement over the last couple of years in IT and it has an impact on integration. No longer can an IT department have full control over the applications that they are integrating. They now have to interact with applications like Salesforce.com.

A number of these trends converged. In the past, you may have been able to address data issues separately with small portion of your IT group within the data center and say application integration separately with another group within the data center. Nowadays, you are not only in control of your own systems, you have to depend on systems that someone else would be supporting for you in the cloud. Thus, the complexity of all of the integration points that need to be managed has exploded.

The architectural trend is really driving the need for the data and application integration technologies and the team supporting those to come together.



These are some of the overriding trends that we are seeing at Talend and responding to in terms of issues that are driving our customer needs today.

Gardner: It sounds like there are two major shifts in addition to some other complexity issues. The two shifts seem to be that we now need to integrate data, applications, and services with some sort of a coordinated effect. Having them in separate silos doesn’t seem to work very well. And then, we have a shift in terms of the architecture of where the computing, the resources, and the data reside -- and that would be this cloud computing activity.

Why is it important for data and application integration activities to become closer or even under the same umbrella?

Walsh: The two trends that you talked about are related. The architectural trend is really driving the need for the data and application integration technologies and the team supporting those to come together. The reason is that data and application integration no longer are necessarily centralized in a single location.

When they were, you had, in essence, a single point of integration that you needed to manage amongst the data and the applications. Nowadays, it’s distributed throughout your enterprise, but also distributed, as I mentioned before, across a network of partners and providers that you may be using.

So many touch points

With that, there’s now the mandate that you can no longer isolate data from application, because the touch points are just so many. You now need to look at solutions that, from the get-go, consider both aspects of the integration problem -- the data aspect and the system and application integration aspect.

Gardner: And, I suppose we need to tool in such a way that we can approach both of these problem sets, the data integration and the applications integration, with a common interface or at least common logic. Is that correct?

Walsh: Yes, and up until now the two audiences have been treated quite differently. I think the tool expectations of the audience for data management versus the audience for application integration were quite different. We're finding that we need to bridge that gap and provide unified tool sets that are appropriate for both the data management user, as well as the application integration user.

Gardner: I think we understand the business requirements now, why this shift is happening, why it’s so important, and how it supports real agility capabilities of an organization. So, this is not a nice to have, but really mission-critical.

Let’s go to Dan Kulp. Tell me why a certain architectural or platform approach best address these issues. It doesn’t sound like a manual, labor-intensive, siloed approach works. Why must we take a different kind of architectural step here, Dan?

Kulp: As Pat mentioned earlier, with the shifting of the requirements from silos into more of a distributed environment, the developers that are doing the application integration and the people doing the data management have to talk a lot more to get these problems solved. Your older solutions, from five years ago or whatever, that had each of those things completely separate were not able to scale up to this distributed type environment.

Gardner: Let me ask you now from a different perspective, architecturally we have a shift, but why does an open source community approach help bring these constituencies together? What is it about an open source and modular approach to these infrastructure components that helps bridge these cultures?

Kulp: One aspect that open source brings is a very wide range of requirements that are placed on these open source projects. That provides a lot of benefit to an organization, as these requirements may not be required of your organization today, but you don’t really know what’s going to happen six months or a year from now.

You may acquire another company or you have to integrate another set of boxes from another area of your organization. The open source projects that you see out there, because of their open-source nature, have been attracting a wide range of developers, a wide range of new requirements and ideas, and very bright people who have really great ideas and thoughts and have made these projects very successful, just from the community nature of open source.

There is also the obvious cost benefit of not having all these high priced licenses, but the real value, in my opinion, is the community that’s behind these projects. It's continuously innovating and continuously providing new solutions for problems you may not even have yet.

Gardner: With cloud computing, you're also dealing with more moving parts. You don’t necessarily know where those parts are coming from or what the underlying heritage is, but if there is an open source commonality among and between them. I'm quite sure that many of the cloud providers have a significant amount of open source in their infrastructure that helps make these interactions, these common denominators technically possible.

New complexities

Walsh: Agreed. The cloud brings a whole new set of complexities and challenges and as you are deploying your applications into the cloud, you need to think about these things. And a lot of these open-source projects that are addressing some of these cloud needs have thought about these things.

If your organization isn’t into cloud yet, but you're thinking about it, leverage the expertise that's already out there. Talk to the communities and get engaged with those communities. You'll learn a lot, and you'll be probably better off for it in the long run.

Gardner: Dan, you've been involved with open source for quite some time in a number of capacities. Maybe you could explain about where you're involved, what sort of projects you are working on, and why this particular mix of projects sort of come to a head in helping us address this integration challenge?

Kulp: I've been involved with open source for roughly six years now, primarily at Apache. I got started at Apache as part of the Apache CXF Project. I've been there since the beginning. As you mentioned earlier, I'm the PMC Chair for that project, very heavily involved.

For those people who aren’t familiar with CXF, that’s the web services stack. At Apache, they're supporting all of your SOAP standards as well as JAX-RS and REST-based services. It’s really a framework for producing services.

As the problems in the enterprise expand from year to year, which they always do, it’s fascinating seeing these open-source projects at Apache being incubated.



Six years ago, that was the problem people were trying to solve. As things have evolved over the last six years, we're seeing more application integration challenges that are beyond SOAP and REST. That’s where projects like Apache Camel come in, where you're doing your enterprise integration patterns inside of your enterprise service buses (ESBs). So, I'm getting more heavily involved with that.

I've also been involved with even things like the Maven Project at Apache, doing build-related tools and deployment scenario things.

As the problems in the enterprise expand from year to year, which they always do, it’s fascinating seeing these open-source projects at Apache being incubated, or even graduating from the incubator, that solve these real world scenarios. To me, it has been an amazing experience to be involved with that whole process of seeing ideas bubble up through the incubator and into Apache projects that solve real world problems.

Gardner: Okay. We understand that there is a new set of requirements for integration. We know that we have an arsenal of approaches vis-à-vis the open-source communities, and some proven and mature projects that are implemented quite robustly in some of the most intensive compute environments.

How do we now bring this together in such a way that your typical enterprise can understand what they can do to bridge this gap between the data and the applications integration and then reduce their risk by setting up an architecture that’s cloud ready or hybrid computing ready?

Let’s go back to Pat Walsh. What are you finding on the street? What are people starting to do in terms of coming to grips with these architectural changes?

Expanded market

Walsh: One interesting point to raise before talking about what we're seeing people doing is that there is an expanded market now for these integration challenges. It used to be that we would see very large enterprises were the ones that were addressing complexity in their organizations.

With cloud-based initiatives and such, it’s affecting even small to medium-size businesses (SMBs). We see a much broader set of enterprises trying to address it. Companies that have fewer than 1,000 employees are now looking at integration solutions to manage their data and their applications in the cloud in a much more sophisticated way than just three years ago. It’s a much broader problem.

The way that people are hoping to address it is by looking for a way that doesn’t require a massive outlay of investment in consulting resources. The traditional large organization, in addition to purchasing product to help them with integrating their data and integrating their applications, would typically have systems integrator help them pull everything together. That’s obviously not an affordable path for an SMB.

Therefore, people are looking to see, how they can find a combined, easy to use way and how they can gain knowledge from people who have experience, having tackled these issues and problems in the past.

We're finding that people are looking for just a simpler, prescriptive way to do the majority of the challenges out there. In terms of the 20 percent outlier problems, you may need to have a systems integrator come in and help you with that. But, people are really focused on the meat and potatoes of the integration of their functions, the data, and the applications that go along with those processes and functions.

We grab those and bring them together, the best of breed from the various Apache projects that solve real world problems.



Gardner: Dan Kulp, we need to have architecture modernization in effect, but we need to do it in such a way that more people in a large organization and more types of organizations, small to medium-sized businesses, can avail themselves of these services, these capabilities.

Tell me a little bit about what you have done to allow that difficult equation to be solved? It seems to me that we are still talking about service-oriented architecture (SOA). In many respects we're talking about ESBs. Five or seven years ago, that was a very complex and costly activity. We've now been able to abstract up the value, but I suppose reduce and subvert the complexity. Tell me how you do that.

Kulp: The first step in that process to solve that problem was identifying where the best solutions are/ They're primarily in open source. I mentioned CXF and Camel, and there is Apache Karaf providing some OSGi stuff.

That was the first step. We grab those and bring them together, the best of breed from the various Apache projects that solve real world problems.

The next step was trying to find or produce a set of tooling that makes using those products a lot easier. One of the things about Apache that you will discover, if you are heavily involved is that we are hardcore developers. For us, writing Java code to solve a problem is natural.

Skill sets

One of the problems that we're trying to address is bringing this great technology produced by the Apache people into the hands of those that don’t have that same level of skill set, expertise, or mindset.

That includes those from the application integration side, where you have developers that are used to doing point-and-click type enterprise integration pattern things, to the data integration people that are used to their data mappings, GUIs, and things like that, and trying to bring both sets of people together into a platform that can solve both teams.

Gardner: A similar questions to you Pat. Where do we bring the value higher but make the complexity less of an issue and less visible? What is it about your tools and approach at Talend that is helping to bring this to the masses in a way that’s automated, a service factory approach, rather than a hand coding approach?

Walsh: Talend has a great history of unifying technologies onto a common platform, to really keep the power of the underlying tools, but simplify the interface to it. This unified platform really consists of five key components.

The first one is a common development environment that is used across the products. The second thing is a common deployment tool that allows you to deploy into a runtime environment.

By providing this unified platform of tools, it allows someone to learn a single interface, regardless of whether it’s at the development stage, the deployment stage, or the management stage.



There's also a common repository that allows you, across the lifecycle of your process, to be able to manage it consistently, regardless of the type of technology that’s being used. Finally, there is common monitoring across the entire environment.

What we are doing now is extending that model that has been applied to our data management products to encompass the ESB, the application integration aspect of it. By providing this unified platform of tools, it allows someone to learn a single interface, regardless of whether it’s at the development stage, the deployment stage, or the management stage, and get the power of master data management technologies, data integration, data quality, or the ESB technologies themselves.

By providing this one interface, this one common environment, allows people to become comfortable with this common interface, but have the benefit of multiple sets of tools.

Gardner: One of the things that I face when I talk about these issues with enterprises is that they like the idea of having more people involved, but they also see that there is a risk involved with that concerning permissions, access, control, and even policy and rules driven activities around who gets to integrate what. How do you solve or ameliorate that problem?

Walsh: We've gone to great lengths to include security mechanisms into the solution, so that we can have approaches whereby there are certain permissions for just individuals. Or, IT management can look at certain aspects while opening it up maybe to a broader audience, when it comes to development and use of the interfaces that are going to be developed on the data in application side.

Democratizing technology

I
t’s very important, as you say, that as we bring this technology to the masses, as we refer to it, democratizing the technology, lowering the barriers to entry that historically have been in place, we don’t remove any of the enterprise qualities that are expected. Security is certainly a major one, as is policy management, so that you could have a number of different business roles that allow you to have the flexibility you need as you deploy it into a large- or even medium-size enterprise.

We're providing both capabilities, simplifying the interface, while not removing any of the enterprise qualities that have come to be expected of the integration products we provide.

Gardner: Okay. Dan has told us a little bit about how some of the open source projects, such as CXF, Camel, and Karaf have provided some fundamental underpinnings for this. But, Talend has also been merging and acquiring. Tell me a little bit about your business and the evolution of Talend that has allowed you to provide this all in one integration capability to, as you say, more of the masses?

Walsh: It came quite naturally from Talend’s perspective. Data customers were using our data integration tools, as well as our data quality tools. We have Talend Open Studio, which is our popular open source data integration technology. Customers naturally were inquiring about how they could provide these data jobs as services, so that they could be reused by other applications, or they were inquiring how they could incorporate our technology into a SOA.

This led Talend to partner with a company called Sopera. They had a very rich ESB-based integration platform for applications. After two years of partnership, we decided it made sense to come together in a stronger way, and Talend acquired Sopera.

We're providing both capabilities, simplifying the interface, while not removing any of the enterprise qualities that have come to be expected of the integration products we provide.



So, we have seen this firsthand from our customers. It really drove us to see the convergence of data and application integration technology, and therefore the acquisition of Sopera’s technology, as well as the people behind that technology, has enabled us to really come in with this common platform that we are just now releasing.

Gardner: The timing sounds very good. There's movement in the market towards democratization, more inclusive platform approach to both data and applications and services integration. The driver in the market about hybrid computing is coming right at the right time in terms of being able to bridge different types of computing environments and integrate across them.

This all is great in theory and we have certainly seen a lot of action in the open source community that had bolstered the ability of these underlying products and projects. But, what about real use case scenarios. Do we have any examples of where this is being used now, perhaps early adopters? Maybe you can name them or maybe you can only describe what they are doing. But for me, showing is always better than just telling. Can we show how this all in one integration capability is actually being used in the field?

Walsh: We have a couple of examples that I can refer to. I think the most tangible one that may make sense to folks is that we have an insurance company that we work with. While they've been working with us for quite some time on the data side of the house, looking at how they can have their back office data shared amongst the different industry consortia that they work with to do ratings and other checks on credit worthiness or insurance risk, that has really been about integrating data on the backend.

Much like any business, they're making it more accessible to their consumers by trying to extend their back-office systems into systems that have more general web interface or maybe an interface at an ATM.

Opened to consumers

So, they required some application integration technology, and with that, they built this web interface and opened it up to consumers. The expectation of their user is a much more rapid response time. When they had to interface with an agent in the office, they may wait 24 hours for a response, but now they expect their answer to come during their web-based session.

The timeframe required has led them to have an application integration solution that can respond in sub-second response rates for their transaction. In the past, they were going with a much longer latency for the completion of transactions.

It's just a typical example that I think folks can appreciate. As people extend their back office systems to consumers, number one, consumer expectations raised the bar in terms of the overall performance of the system, and thus the technology that’s supporting those systems needs to necessarily change to support that expectation.

Gardner: In listening to Pat describe that use case, Dan, it sounds as if what we're trying to accomplish here is to do what the data warehousing, data mining, and business intelligence (BI) field have done, but perhaps allow many of those values to be extracted with more agility, faster, and then with a dynamic approach.

Is that fair? Are we really compressing or creating a category separate from BI, but that does a lot of what BI does vis-à-vis the integration of data and activities for application services?

That requires a whole new set of skills, a whole new set of challenges.



Kulp: That’s exactly what’s happening. A couple of years, data mining ended up being batch jobs that were run at midnight or overnight. Then, the data would be available to the front end people the next morning. You'd get your reports or you'd log into your system and check the results of these batch jobs.

With extending your backend data systems to the consumer, these overnight batch systems are really not meeting the expectations of the consumers. They're demanding that their information be available immediately. They submit a new request and they want to have things updated immediately, so that results are available and displayed within seconds, not overnight.

That requires a whole new set of skills, a whole new set of challenges. The people that were doing the front-end application integration that queried the data from the overnight batch jobs suddenly have to have some expertise in not just cleaning the data, but allowing or working with the team doing the data space, to provide updates to that information in a much more dynamic form.

Gardner: How is this going to become more critical? Looking to the future, particularly for organizations that are doing more and more web-based commerce, perhaps even more mobile commerce, whether it’s through a web interface, a HTML5 interface, native applications on mobile devices, it seems to me that the consumer activities are driving more need for this fast feedback loop integration and data analysis function.

Let’s start with you, Pat. Why in the future does what we are talking about today become even more important, therefore become more critical as a core competency?

Becoming more relevant

Walsh: You can see that, as the consumerization of technology increases. We're already seeing the pressure that IT feels from becoming more relevant to the business, that just expands.

As I said before about the consumerization of devices in the workplace, it really does come down to the interfaces and the expectations that it doesn’t require a specialist in an IT field to be able to manipulate and analyze the information that they need or even to create a service or application that would enable them to do their everyday task or work function.

That’s just going to expand it. It has been happening, and we are just going to see that at a more rapid pace. It’s going to require that vendors and technology companies like Talend respond in kind and build products that are more accessible to a broader audience of users.

I think it’s analogous to what we saw in the early days of the Internet. Early on you would do command-line interfaces to send files back and forth. Once there was a web-based interface, it opened it to the masses. Nowadays, we think nothing of using a web browser to do all kinds of activity that 20 years ago was reserved to just people that had a technical know how to manipulate those systems.

We are seeing the same across these aspects of the business that up until now had really been the bastions of IT teams.

If it’s beneficial to my organization, why wouldn’t it be beneficial to others in my industry or to an even broader audience?



Gardner: I would also wonder if data services become additional revenue sources for companies. If they can expose just the right amount of data safely and securely and give people some tools to work with that, not only do they provide services, but the fact that they were in a position to gather data about certain markets, certain activities, be it B2B or B2C, they can then in a sense monetize that data back out into a field of partners and/or end-users.

Is there an opportunity for enterprises to start looking at data, not just as an asset, but as actually a product or service to sell?

Walsh: Absolutely. Today, we see that they are really addressing data services as an efficiency within their organization. How can I leverage the investment that I have made in this initial data analysis or data job across the entirety of my organization? But it’s not a big step to take beyond that to say, if it’s beneficial to my organization, why wouldn’t it be beneficial to others in my industry or to an even broader audience?

So we absolutely see that as a level of commerce that will be enabled by more sophisticated data services, technology, with a more accessible interface to that technology.

Gardner: Dan Kulp, same perspective of the future of what’s going on in the future to you. How do you see the trends around mobile and even localization services and mobile commerce? How do these shape up, so that we will require more of the types of services we have been talking about today, that all in one integration, rapid iterative development around it?

Comes down to consumers

Kulp: It really comes down to the consumers of these services and data. As the markets have expanded and the consumers are demanding things to get their information faster or get more information or advertisers need to figure out, where are these consumers going and just the whole variety of information sources expand out as well, the architecture of the applications and the interactions between the front end and backend systems kind of get blurred.

Things are changing, and companies like Talend that are involved in the space need to adapt as well and provide better solutions that make these blurring lines occur a lot quicker. That’s what we are trying to target today.

Gardner: We will have to wrap up now. We're about out of time. Pat, for those folks interested in learning more, do you have some resources, some white papers, reports? Where would I go if I wanted to learn more about this integration across data and applications function for the masses? What do you have available?

Walsh: The easiest place to go would be our website at www.talend.com.

Gardner: Dan Kulp, what about in the open source community? Can you point folks to a place where they can learn more about some of these underlying and supporting projects?

The community behind those projects is as much of an asset to the projects as the code itself.



Kulp: Each of the projects have their own website with information. So CXF is cxf.apache.org; Camel is camel.apache.org; Karaf is karaf.apache.org. However, if you just go to the Apache website, at www.apache.org, there are links to all of them, as well as a lot of valuable information about how Apache works and how these Apache communities work and how you get involved?

A lot of that is just as important as what the technology projects themselves are trying to solve, but the community behind those projects is as much of an asset to the projects as the code itself. I encourage people to poke around there and see all the exciting things that are going on at Apache.

Gardner: You've been listening to a sponsored BriefingsDirect podcast discussion on how the role and impact of integration has shifted and how a more comprehensive and managed approach to integration is helping enterprises produce and leverage more data driven business processes.

I'd like to thank our guests. We've been here today with Dan Kulp. He is Vice President of Open Source Development at Talend’s Application Integration Unit. Thanks so much, Dan.

Kulp: Thank you.

Gardner: And also Pat Walsh, Vice President of Marketing at Talend in their Application Integration Division. Thank you, sir.

Walsh: Thanks, Dana.

Gardner: This is Dana Gardner, Principal Analyst at Interarbor Solutions. Thanks for listening, and come back next time.

Listen to the podcast. Find it on iTunes/iPod and Podcast.com. Download the transcript. Sponsor: Talend.

Transcript of a sponsored BriefingsDirect podcast on enterprise integration and new tools to put control in the hands of "the masses." Copyright Interarbor Solutions, LLC, 2005-2011. All rights reserved.

You may also be interested in: