The ITKO Blog: Service Virtualization and Application Lifecycle Optimization - CA Technologies

Dan Woods on Brick Walls Of Virtualization

Brick Walls Of Virtualization is the title of a recent post by Dan Woods at Forbes.com. The subtitle is, "Without automation, creating and managing system images will slow adoption." I could not agree more.

The "brick walls" refer to the burdens of virtualization that come along with the benefits. On the benefits side Dan mentions the cost savings of less physical servers; increased flexibility to provision systems or change configurations; and better access to resources in the cloud. This latter benefit outsources complexity and shifts capital expenditures to operational expenditures.

You would be hard-pressed to find a large company not already leveraging some mix of hardware and software virtualization. Especially in a down economy, it has obvious value.

However, Dan then brings up the burdens of brick walls. The good news is that the number of physical servers may drop due to support cost and energy pressures. And, the potential bad news: the number of virtualized servers and devices can skyrocket without the physical constraints, and each of these virtualized devices still needs to be managed. We are seeing this all of the time in our major enterprise client engagements - virtual instances can proliferate quite rapidly if not managed carefully.

Here's another, deeper challenge of virtualization in enterprise IT: for all of those virtual servers, you still have dependencies or "wires hanging off" of those, leading to big iron mainframes, SaaS applications, 3rd party services, what have you. These dependencies are resistant to hardware virtualization, so they need to be simulated with behavioral virtualization.

I personally like to compare this phenomenon to unmarked leftovers in a shared refrigerator... [whose food is this, and without opening it up, can I figure out how long it has been here?]

Dan adds that a "whole new class of IT systems-management software and techniques is needed to create a virtual data center operating system." Part of the increased requirements for systems development and administration involves testing and continuous validation, and part of this involves better management of both VMs and Virtual Services, a topic we will discuss in further detail on this blog.

There is much more in Dan's post and I recommend going to Forbes.com to read it. I liked his concluding remark, "The upshot is that IT, which is dedicated to automating the rest of the business, must start spending more time automating itself." Well said.

Realizing the Benefits of Enterprise Integration and SOA at Lower Costs

We have now made available the complete podcast of a recent webinar in a compact video: Realizing the Benefits of Enterprise Integration and SOA at Lower Costs, featuring iTKO’s John Michelsen and Perficient's Eric Roch who cover how to achieve shorter release cycles and lower costs within SOA implementations. Here's an overview of some of the topics covered in this well-attended session on iDevNews.

Eric begins with key challenges to successful EAI and SOA implementations and some case study examples. Perficient is involved in technology implementation with over 1500 consultants. Eric leads the Business Integration Practice that includes BPM and SOA. He covered integration testing that includes a combination of services, components, and systems. Eric said that the scope of most projects causes test environment complexity issues and service availability issues as we have often discussed here. Often quality assurance is pushed to the end, and companies end up with infrastructures that do not perform. There are also complex dependencies between development groups.

Tools are also often dated holdovers from client server days. So companies must either build or buy new test tools. Most companies start off in a build mode but they end up with tools or custom "test harnesses" that cannot be used in a repeatable way. The test harness building option can become more time consuming than building the application itself. Eric encourages companies to make a well-considered test tool decision to ensure there is sufficient agility. Eric then went into great detail on the specific test requirements in SOA and provides a case study from an electric utility company where they employed multi-tier testing with LISA, and used the Virtual Service Environment capability to simulate "test reponsive" endpoints on the ESB, Services and data layers, rather than hand-coding their own responders. See the attached diagram for an example:

iTKO founder and "Chief Geek" John Michelsen then drilled down into a bit more detail on testing, validation, and virtualization within SOA. He focused on areas where you can drive out costs and provide greater efficiencies. John mentioned that traditional testing methods cannot automate testing in the complex world of SOA and this limitation produces significant inefficiencies. We can now build applications faster than we can test them using the traditional methods, and this adds unintended costs to SOA implementations. The power of SOA to integrate services raises the complexity to test this integration.

John covered how iTKO LISA automates testing, provides continuous validation of the integrated services, and offers virtualization to get around availability constraints on testing and validation. Virtualization gives almost infinite ability to virtualize constraining services and also strips out up to 90% of traction costs for these services. He mentioned an example from a cable company where a simple code change caused significant ramifications to the services that were affected. Continuous validation can pick this up.

Then a Q&A session allowed for further exploration and a great deal of interaction. The first question asked if the John and Eric had noticed project failures when firms attempted to integrate several ESBs, often as the result of mergers or acquisitions. John gave a number of examples and how the issues were addressed with iTKO after the problems occurred. Eric offered other examples and how they handled the issues.

The next question asked how to use virtual environments for integration testing. When can you use service virtualization versus requiring a live validation? John aid that this is a great question as it gets to the heart of the iTKO virtualization approach. LISA does not virtualize you, it virtualizes the things you are dependent on. In about 80% of your development cycles, that fact that your downstream service is live is a problem that service virtualization can address. Now in the final stages of integration testing, then you want to test against live data. So you virtualize your dependencies to take costs out during development and then switch to live data at the end. Eric offered some more examples.

The next question asked if John and Eric had experience with the virtual validation of SOA governance and how this worked? John first replied. LISA provides the infrastructure to do this and experts at firms like Perficient can use LISA and integrate it as the validation engine. Eric added that governance is about people, processes, and policies. He likes to fit the governance model to align with the culture of the organization. Then you measure it and change it over time. You cannot get too heavy handed about governance without creating adoption problems. Eric went on to offer more details on how they do this.

The next question asked how a delivery partner like Perficient can keep its own costs down in a complex implementation to stay competitive. Eric said they bring expertise and they have a set of methodology to increase project velocity. He often sees that SOA is too focused on architecture and services but not covering the business issues and the quality concerns. It is about the process and not the technology.

Then the moderator asked why testing is such a big chunk of SOA implementation costs when there is now a modular architecture. John said that at the granular level testing is easier. The problem comes with the overall number of dynamic parts and the resulting complexity. Eric said the problem of dependencies in the lifecycle has existed for a while, even with mainframe systems. You can get efficiencies if you can eliminate dependencies.

We encourage you to hear the complete podcast. It is available at our iTKO Resources site.

Showing some ERP mettle for SOA with SAP

So many of our customers have large-scale ERP packages like SAP or Oracle that form the IT foundation of the enterprise.  While at the outset this sounds like a simple architecture, given that you are working with a single vendor over a longer timeframe, in fact these systems are themselves rife with complexity and heterogeneity.  And so many applications now connect into the ERP via various forms of adapters and message buses, that you end up with the classic alphabet soup of technologies at play.

"iTKO/SAP Whitepaper: Validation & Virtualization of SAP Enterprise SOA Applications with iTKO LISA"

This new paper iTKO co-authored with SAP, after working with them in their labs, helps customers see just how valuable a service validation and virtualization offering can be in these environments.  By SAP’s own calculations, customers spend millions of dollars every year standing up and maintaining development and test labs for SAP, even when running on virtual hardware.  And when they need to perform complex integration and load tests with other systems, millions can become $10s of millions to support these environments.  

Our strategy of virtualizing the services layers of these complex applications in a VSE (Virtual Service Environment) like LISA offers is a great way to eliminate those costly constraints, and make it practically zero cost to stand up. This approach is great for our software partners like SAP, and especially for the end customer. After all, we all want to focus IT budgets and especially our integration and innovation efforts toward technologies that contribute revenue to the bottom line, and not manual support and configuration costs that aren't repeatable.

Yes, even big ERP is getting nimble for SOA. We expect even more collaboration with SAP and other software partners to better improve quality and agility for customers with Virtualized Services. You can also find the paper posted by co-author Joerg Nalik Ph.D. on SAP's Developer Network and COIL blog here: https://www.sdn.sap.com/irj/scn/weblogs?blog=/pub/wlg/14083.

Cloud Insanity - Stop Repeating Mistakes of the Past

I enjoyed Mike Kavis’s post, Cloud Insanity - Focusing on all the wrong things, again, especially his definition: "Cloud Insanity - To repeat the same behaviors that caused SOA to fail and expect a different result." How true. We have written a good bit on this blog about how SOA can succeed if you pay attention to the right stuff, like the governance issues, the business issues over the tech issues, and change management (see for example, Avoiding SOA Disillusionment - eBizQ Panel Recap).

These are some of the issues that Mike also lists as things to address with both SOA and the cloud. He goes on to say that we should focus the cloud debate on how to "tackle security and compliance, how to manage hybrid solutions, how to provide reliability and scalability, how to prevent latency, and many other important topics. But enough of the Cloud Insanity."

As with any hot technology, we have many vendors and consultants repackaging their expertise as a "miracle tonic" for Cloud Computing, without regard for the drivers that determine where it makes the most sense for business. We should apply the lessons learned from successful SOA efforts to help ensure that cloud efforts are successful (see Avoiding the Hidden Costs of Cloud Computing with Lessons Learned from SOA).

One of the additional lessons similar to service-oriented development is the need for systematic testing that makes use of automation and virtualization as much as possible, so we can validate changing cloud-based services as part of a continuous business process. In return, Cloud computing is a boon for test labs and virtualized service environments, adding a lot of flexibility and options in how these assets are deployed for the business.

Comparing the State of Consumer Web and Enterprise Applications

Here is an interesting post by Jeremy Chone in Bits and Buzz on Enterprise Web vs Consumer Web [2.0]: Top Six Differences. I was struck by this comment on one differences and thought of similar things we've seen in the field. Especially the fifth factor.  Here it is in Jeremy’s words.

"5) Integration (Loose vs Strict) ... In the consumer space, application integration happens very organically. Application providers (e.g., Twitter) or application container providers (i.e., Facebook) define  interfaces with which everybody can integrate. The result is that thousands if not hundreds of thousands of applications will potentially integrate with some of these interfaces. Consequently, interfaces tend to be simpler, lighter (e.g., REST), and often even client-centric (e.g., OpenSocial and iGoogle). In other words, consumer application integration is about breadth rather than depth. In the enterprise space, on the other hand, application integration has to be thoughtfully designed and managed and often has an impact in all layers of the application (i.e., presentation, logic, and data). Heavy and expensive technology, such as ETL, SOA, JMS, and Portal, are often an effective way to establish strong integration between applications."

Many of the current enterprise application efforts have been influenced by the consumer web (and you can call it 3rd party services, Cloud, SaaS, what have you).  However, success depends on recognizing the challenges and differences, as much as the opportunities and similarities. Much of the consumer web involves preparing for the unknown such as traffic volumes, as well as integration from multiple and unanticipated applications that you may not have control over, or be able to enforce an expected service level on.

While more is often known about the architecture in the enterprise world, the level complexity can go way up due to much deeper layers of system-to-system interaction.  There is also less tolerance for lapses in performance and data integrity.  This is where the edge between Consumer Web and enterprise applications gets particularly jagged, and where we are seeing the most innovation.

For instance, the need for software testing to be truly stateful and asynchronous as you cross over the edge between the enterprise and consumer web is a necessary thing (and quite a mouthful) -- if we are going to consume the best of both in our distributed apps.

How do you validate a transaction, if the response may move around among several third parties and decision points, and maintain its state when (and if) an appropriate response occurs? This kind of leap is going to take a lot of automation and innovation to allow the two to achieve scale.

Gartner Sees Expanding Role for SOA

I like the way this post, Gartner Says SOA Is Evolving Beyond Its Traditional Roots, starts out: "Despite the economic downturn, Gartner analysts said they do not anticipate a dramatic drop in SOA adoption."

It goes on to quote Jess Thompson, a Research VP we've enjoyed talking to at Gartner: 
"SOA underpins and enables initiatives aimed at cost reduction, such as SaaS or BPM, and we have noticed a growing interest around SOA from vertical industries like government and healthcare, which have been waiting for the technology and best practices to mature."

The post goes on to acknowledge that there are horror stories about SOA projects failing that have created a perception that the era of SOA is over. However, they add that the lack proper SOA governance is, and will be through 2010, the most common reason for SOA failures. This is something we have discussed a bit on this blog (see Avoiding SOA Disillusionment).

They conclude that quite contrary to going away SOA is actually emerging from the Trough of Disillusionment within Gartner's Hype Cycle and climbing the Slope of Enlightenment. It is good to see this trend in the field. Indeed, their previous commentary on SOA talked about fewer enterprises starting new SOA projects in 2008 - which is true - when you consider that any expansive IT department is already leveraging elements of an SOA model and it is quietly being used to create value.

Updated iTKO Website - 3 minute intro, new solutions areas

A quick note: We’ve made some significant improvements to the iTKO website (www.itko.com) today that will provide clearer starting points for learning more about testing, validating and virtualizing today's distributed enterprise apps.

As a highlight, look under Next Steps, where there is a cool new 3-min Flash movie that summarizes the benefits and capabilities of LISA in the context of the 4 C’s. Take a look:  http://www.itko.com/site/demo.html

We’ve also expanded the Solutions section to covers several new categories including ALM, cloud computing, outsource optimization, and others. We think the updated UI is a little cleaner and hope you agree. Let us know what you think!

 

Meeting the Business User’s Expectations with Cloud computing, SaaS

Neil Ward-Dutton recently posted a comprehensive piece on Cloud computing, SaaS and SOA - the universal service network. He wrote that "Regardless of where services come from (and indeed because they will come from multiple places, creating cross-enterprise service networks), it's increasingly the case that in order to deliver effective IT capabilities in the 21st Century, you need to understand SOA principles and build technology and management structures that really support the principles of service orientation."

He then goes on to discuss two sets of business user expectations. First they expect flexibility. Neil writes, "when business teams that are using SaaS-based offerings learn about the infrastructure side of the story - how easy they can be to customise, extend, and integrate with - and ask why internally-developed systems don't exhibit the same qualities? Another slab of SOA pressure, that's what." The business will demand a high level of flexibility and control in how service-based offerings are composed and consumed.

Later Neil brings up the second expectation. He refers to Todd Biske, on ITIL and SOA, and states how it's "crucial to consider the full service lifecycle when you do SOA, and drive governance to ensure that you can deliver "real service" - not just code wrapped in XML-based interfaces." Then he offers some examples where the provider's idea of service doesn't match the consumer's expectation.

In addition to flexibility, businesses want applications that work. These two points go to the heart of the need for higher levels of test automation and agility throughout the design and development process, as well as the benefits that virtualization can bring to both flexibility and getting good, ready data.

Without rapid, repeatable test cycles, you cannot achieve flexibility as development cycles get bogged down. This was the fate of one financial service institution before they moved from manual to automated testing. Service Virtualization can further accelerate the development process by eliminating the bottlenecks of constrained or unavailable systems as one European telco discovered.

Neil concludes that the "rise of cloud computing and SaaS should entice us to revisit our development-centred assumptions about SOA and search for a 'bigger picture' that focuses on consumer expectation and value first." One way to do this is to simplify the development process and enable faster iterations. This enables us to focus less on the integration challenges of the technology, and more on the business challenges the technology is designed to meet.

Service Virtualization Addresses the Sticky Challenges

Alex Handy recently wrote about virtualization in testing for the San Diego Times. (see Virtualization bridges, from development to IT). He notes that virtualization is a hot IT topic and begins with the  comment that, “many of virtualization's proponents see the technology as a way to bridge the gap between operations and development. The path to that bridge, they say, is often through the test lab.” He later notes that on the application side, developers are motivated by speed and flexibility. These concepts are often completely lost on the operations side, where the focus in on cost and control. Virtualization addresses both needs.

In addition to saving time and money, Alex notes that it can provide continuity to a support team as it can look at a given user's exact deployment of a given application or a given group's exact configuration. This can really accelerate the support and remediation process.

One challenge arises when virtual test beds attempt to enter distributed application environments, such as Cloud Computing and SOA. When an application is tied into a network of services offered throughout an organization's extended network of partners, how can they be added into a test environment? You can't just take a hard drive or a server and replicate it, when many of the elements in the environment aren't available for that activity, or they are far too complex to realize cost-efficient hardware virtualization models.

Here Alex quotes iTKO's John Michelsen, who said that this issue can be troublesome for large development projects inside and outside of the cloud. "In the real world, there's an airline that has 1,500-plus regression tests that run on a nightly build," said John. "They spent three or four hours each day dealing with about 10% of those tests failing every night. The vast majority of the failures were false failures." The lack of availability to the needed services, and relevant data for testing, made the test lab a major bottleneck during the software lifecycle.

Service Virtualization, using a Virtual Service Environment to capture and simulate the behavior and data of interconnected systems, addresses this issue through bypassing the constraints of operating in the real world with production systems. Here is an example of how an airline addressed this issue with virtualization. 

Avoiding the Hidden Costs of Cloud Computing with Lessons Learned from SOA

George Lawton recently provided a post describing how "as with SOA, some development costs obscured by cloud computing." He extensively quotes iTKO’s John Michelsen and this is greatly appreciated.

George writes that while cloud computing holds great promise, it will be filled with surprises, and some of these will reflect surprises encountered in the early going of SOA. The more loosely coupled and distributed the cloud based system becomes, the more moving parts you don't have control over -- and the more hidden development costs lurking within this loose coupling.

Much of cloud functionality is created and hosted outside the organization, which creates ideal "pay as you go, leave when you want" cost and agility advantages, but it still requires testing, perhaps more than ever. To compound the problem "after the application is deployed, the service provider can make changes that could adversely affect the application without notice."

These needs call for automated testing and continuous validation. Then new two needs emerge as George notes. "After an application has been tested and deployed, paying per-unit of service (storage, CPU time, and bandwidth) allows a new app to get deployed with minimal cost. But these fixed costs can take a hidden toll when a company has to do any sort of testing, and particularly load and performance testing of a new app."

For instance, take a firm we know, whose operations team needed to run a performance test to verify required Service Level Agreements (SLAs) to verify that customer response times would be within bounds. One routine test cost about $10K in access fees from the service provider - an unforeseen bill that wouldn't have appeared if the team was banging on their own servers. That inhibits the team from testing and adjusting their service appropriately.

In addition, there is lack of control and ready access over many of these external services, especially if they are still being developed. This is where service virtualization comes in. Now you can bypass costs and areas where you cannot easily set up test situations on dependent services. As George notes, the methodologies and tools that were developed for SOA testing and validation will come in handy to handle cloud applications.

Of course, none of these challenges mean Cloud-based development isn't a great thing. We love it for all the reasons we loved SOA, then SaaS-based models, because it gives the enterprise incredible flexibility and lower cost of entry for needed infrastructure and applications. We are also seeing a new level of quality by pushing Virtual Services to provisioned Cloud environments for testing and performance labs, in order to keep unexpected third-party costs under control and allow unfettered access to test and development teams.

Expect to hear much more on this front as Cloud Computing gets tapped for serious, distributed enterprise applications that need validation and virtualization.

James Governor on Combining Cost Reduction with Innovation Through SOA

James Governor, co-founder of RedMonk, recently spoke at IBM Virtual Global SOA Forum. Joe McKendrick provides a transcript of his talk on ebizQ, "Work Smarter, Take Out Costs In a Tough Economy." He begins with saying that he "wants to talk about culture, because service oriented architecture clearly sounds like something technical, but really if you want to succeed, its going to be all about the people."

He goes on to say that SOA is the language of change. SOA was started, in part, to enable businesses to become more flexible. "Enterprise customers were concerned that their cost of integration was very high. Once they had built a system, effectively, they poured concrete onto it." This probably affected the thinking about business process and the ability to innovate, and not simply the technology components.

In today's economy, change is needed more than ever. It now takes on two components, the ability to quickly adapt to changing market conditions, and the ability to do this in a cost effective manner. Though SOA was once an overhyped term, there is a "little soa" in the way companies are now quietly using service-oriented approaches to create real value.

"Little soa" encompasses everything else lumped under flexibly leveraging technology to meed business needs - from BPM to integration platforms, rich Internet apps and REST-based approaches that have business services under the covers. "Little soa" can address both of these components and hopefully changes the way people think about innovation and the capacity to engage in it and implement it. This is where the people side of technology comes in.

James goes on to say, “We're going along this road, and it’s very important that we need frankly, a fast car that will keep us on the road. And SOA is a language that the business can talk to IT and systems, in order to stay on that road. In order to have alignment, so that we can move forward, and actually deal with change, and deal with these curves that we're being thrown."

He later mentions virtualization. This is one way to keep the fast car moving forward and stay on the road at lower cost. We have been covering this in our series on Field Reports on SOA Success. See for example, Part Five: Major Airlines. Here the airline was able to virtualize the kinds of services that are normally prohibitively costly and complex to replicate using conventional means, and keep their SOA development and testing on track.

(Concrete Photo by Brenda Anderson and Car Lights Photo by miss blackbutterfly both from Flickr.)

iTKO LISA 4.6 Released

Ok- this post is about ourselves - but we do have cause for celebration. We are thrilled to announce the release of LISA 4.6 Suite today. We believe this latest edition takes LISA miles ahead in terms of Service Virtualization capabilities and breadth of support for heterogeneous technologies, rich internet UIs, and ease of use.

You can read the official iTKO press release here:

iTKO Releases LISA 4.6 Product Suite with Third Generation Service Virtualization Capabilities, Enhanced UI Testing, and Expanded Policy Validation and Governance Support
LISA 4.6 Suite Proven at Key Enterprise Customers and Reviewed by Analyst Firm Butler Group

Congratulations to the LISA development team, and many thanks especially to iTKO's beta customers who stepped up to validate this release and provided us with insightful feedback over the last months.

Additional coverage on LISA 4.6:

• Infoworld: iTKO touts Web 2.0, virtualization in test product

Butler Group Conducts Independent Audit of iTKO

The Butler Group, a UK-based IT analyst firm, recently completed a third-party technology audit on the LISA product suite v4.6. The audit is fairly broad as it covers product functionality, operation, and deployment along with commentary on product strategy and future directions.

Butler Group technology audits are free to their members, however we paid a fee enabling us to have limited distribution rights. You can download a copy on our site: see - Butler Group View - LISA 4.6 Technology Audit. Although we paid a fee, Butler Group performed the audit objectively and authored the content.

Butler Group is the IT End User division of Datamonitor and is a leading provider of Information Technology research, analysis, and advice. Founded in 1990, Butler Group is respected for the impartiality and incisiveness of its research and opinion. We are pleased to offer this report. Here is the abstract:

The iTKO LISA Suite, comprising LISA Test, LISA Validate, and LISA Virtualize, provides capabilities around testing, validation, and simulation of services in a distributed, multi-tier, component-based infrastructure. The code-free nature of the solution enables QA teams and developers to easily conduct unit, functional, regression, and load testing in test environments consisting of either live deployed services, or services that are unavailable or yet to be built. These tests can be performed in a simulated virtual environment that mirrors the behaviors and capabilities of an actual production environment. LISA is capable of running tests and tracking metrics at multiple underlying technology layers (for example, the Web UI level, data integration level, ESB messaging layer, Java objects/EJB and so on). The solution is targeted at Global 2000 organizations with significant investments in SOA, BPM, integration suites, ESBs, or any component-based architecture, looking to streamline their test and development processes for composite applications.

If you would like to see the entire report, it is available at Butler Group View - LISA 4.6 Technology Audit.

Darwin’s Take on Application Developers

Charles Darwin was born 200 years ago on February 12. It took a while for his ideas to take effect, but they are pervasive now. Darwin wrote way before computers but as Mike Gualtieri of Forrester recently wrote, his ideas on "survival of the fittest is an elegant explanation of why so many species exist, why some become extinct, and why some flourish."

Mike recently reflected on what Darwin might say about application developers. He wrote that, "as long as developers continue to write billions of lines of code, they are not going the way of the dinosaurs or the buggy whip manufacturers anytime soon." However, there is a caveat to this. Mike quoted Darwin on the secret to application developers' success. "It is not the strongest of species that survives, nor the most intelligent, but rather the one most responsive to change."

IT is a field of constant change so successful application developers are those that follow Darwin's pronouncements. Agile and SOA are two of those changes in paradigms that require nimble thinking. Success involves moving off of old ways to fully embrace all aspects of the new approaches. A (belated) Happy birthday Charles Darwin. (Photo above by ChrisL_AK under Creative Commons).