By: Jan Stafford
In the United Kingdom, applying for university and other higher education courses is quite different than in the U.S. In the U.S., students apply to individual institutions and get responses one by one over several weeks. In the U.K., all applications are approved online in a single day. Imagine the load placed on a web service that doles out all a nation’s higher-education admissions information in one day. That demand lead to the creation of a scalable, secure hybrid cloud solution on Amazon Web Services (AWS).
“Every over-18-year-old student in the UK who has applied to a university tries to get on the website on that day to see if they have their university place,” said Steve Millidge, director of C2B2, a Malvern, U.K.-based consulting firm. That’s about 650,000 students who’ve applied to hundreds of institutions. “The workload is incredibly thirsty,” he said.
Millidge’s C2B2 development team helped the major UK higher education software provider optimize a college admissions and application management service and move it to the cloud. Millidge has used AWS since its early days, first to test middleware products and then as a platform for quickly building up and tearing down environments. These days, he’s building and testing production environments in AWS for universities, government agencies and other businesses. Millidge is also a resident expert for SearchAWS.com.
The UK service provider used to run the admissions site on an on-premise system, and service delays and outages were a constant headache. Last year, the provider moved to a hybrid cloud approach, supporting the admission applications in its on-premise data center and in an AWS environment. No downtime, real-time analytics, high performance and scalability were the key desired results.
During this project, C2B2 helped the provider move its Oracle Fusion middleware into the AWS platform and build the technical architecture that would support this traffic peak. “We helped build out a replica of the in-house environment to manage specific events and integrate the data between the two environments,” said Millidge.
C2B2’s middleware expertise helped the provider avoid common, but incorrect, assumptions about the networks and the availability of shared storage, local services and servers that the service would run on, according to Millidge’s colleague, Nick Wright, C2B2 senior consultant. These were quite reasonable conventions for the time. Assuming everything is located in a single data center, these traditional systems can be used to create very effective architectures. Over time, however, the legacy components deployed and the interdependent configuration will not scale or be flexible enough for rapid changes common to cloud environments.
Moving away from brittle interdependent services takes a lot of courage on the service provider’s side, Wright said. “We were having to break the conventions that were being relied upon by legacy middleware components.”
C2B2 took a staged approach to solving the problem, identifying pathways that the applications used and analysing how to provide business functionality by isolating micro-services used in the workflow. “We then looked at various ways of bundling these services together so that each set of dependencies can be managed in isolation,” said Wright.
An important step was building high availability and automation into the services, so that they could be treated as a single surface that we could scale on demand. “This final part is where being in the cloud really helped,” said Wright. “We were able to leverage cloud automation tools as an API (application program interface) to translate our service needs into provisioned infrastructure.”
After deployment on a scalable, secure, AWS hybrid cloud last year, the provider ran the university admission process without downtime or operational incidents. All the while, the service provided real time feedback to end users on application performance. As a result, the provider confirmed the placement of 385,910 students into higher education on the A-level results day.
Jan Stafford plans and oversees strategy and operations for TechTarget’s Application Development Media Group. She has covered the computer industry for the last 20-plus years, writing about everything from personal computers to operating systems to server virtualization to application development. E-mail her at email@example.com.
There isn’t a proven one-size-fits-all method for selecting a software tool. Many IT professionals recommend devising a list of necessary features and a thorough evaluation of product offerings, but that’s just not Jake Kooper’s style.
Kooper, who develops mobile applications as a side project for Whitewater Labs, takes a different approach. “I don’t sit down with a checklist and go one, two three, this is what needs to happen,” he said. “I don’t read anything on their [vendor] website. I just sign up and see what happens.”
While Kooper admits his tool selection method may not be appropriate for everyone, it’s what works for him. As a developer, he said he learns best by taking a hands-on approach, rather than looking at infographics and reading lengthy documents.
By visiting a vendor’s website and downloading samples, Kooper is able to start building an application right away and get a feel for how the tool really works. When he runs into a problem, that is when he turns to documentation.
What defines a good tool is quite simple in Kooper’s mind. “At the end of the day, you want to just make sure the product does what you need it to do,” he said.
How do you go about selecting a tool? Are you the type to follow a stringent plan or do you just wing it? Are there any special tips or tricks you use?
Nearly six months after the release, the jury seems to still be out as to how, or even if, UML 2.5 is doing what it’s intended to. The goal of the update wasn’t to change the language, but to simplify documentation and make life easier for developers.
Ed Seidewitz, chief technology officer for the Americas, said many people outside of the Object Management Group (OMG, which adopted UML in the 1990s) are just starting to dapple with the latest UML version. “UML 2.4 is still the most prominent out there,” he said, noting that feedback has been positive from those who have begun working with UML 2.5.
What users see and use UML for didn’t change for the most part, notes Cory Casanave, CEO of Model Driven Solutions, but it “really puts a maturity level into UML.”
A big gripe with UML among developers was that tools failed to work well with one another and models didn’t transfer well. One of the main goals of UML was to improve interoperability between tools and models. “It’s too early to know if that will have a significant effect in practice because most vendors have not implemented UML 2.5 yet,” said Steve Cook, Architect at Microsoft.
That being said, from what Casanave has seen so far, “UML 2.5 has largely made that problem a thing of the past.”
Have you started working with UML 2.5? What do you think of the revision?
By: Jan Stafford
When building dependability into systems and software, the buck often stops at the enterprise architect. While the architect plays a key role, ensuring dependability is a team effort, and business executives can and often should be project leaders, said Chris Forde, The Open Group’s general manager, Asia Pacific, and vice president of Enterprise Architecture.
Forde offers advice on achieving system dependability and career opportunities in embedded software development in this Q&A. Before joining The Open Group, Forde was an enterprise architect for several organizations, most recently as vice president for Strategy and Architecture for their Customer Servicing Capability for American Express. The Open Group is a vendor- and technology-neutral advisory, standards-creation and certification organization, which recently released the Dependability through Assuredness (O-DA) framework. For Open Group, Forde played a key role in the development and launch of TOGAF 9.
What are some indications that a system is not dependable, other than just shut downs?
Forde: The frequency of degradation in service levels observable by the customer (internal or external) from nominal, not failures. This is often cited as more frustrating than total outages, it may also go unreported.
Considering that there is no 100% dependable system, what are the practices that will lead to a low failure rate?
Forde: One is the actual prevention of the failure by recognizing the issue in its early stages as it is developing and taking corrective action in advance; that is, being able to measure prevented failure. A simple example is in effective management of storage or network capacity. Fundamentally, capacity planning and timely provisioning of certain infrastructure components [reduce failures]. Often, this is not done well.
How does constant change impact dependability?
Forde: Change introduces the risk of error, complexity and interoperability compound that risk. O-DA is a mechanism to address mitigating that risk.
What kind of challenges does a mix of cloud, embedded, web and mobile pose for an enterprise architect?
Forde: To a certain extent, architecture implies the effective use of known standards/component or in some cases developing and setting those standards within an enterprise. As these technologies emerge and evolve, they cross into the ‘extended enterprise space’ and reflect the exponential proliferation of endpoints and platforms and information. A lack of standards around emerging technologies is in itself a headache for architects. Attempting to construct fit for purpose solutions with a reasonable lifespan to in order recoup investment, and allow a migration path as things change is a common issue concerning architects.
In the embedded systems area, what career opportunities do you see for enterprise architects today?
Forde: The fact that there are embedded systems is related to expertise, if these systems have a position in an enterprise relative to other systems then they are in scope for an architect’s role. For [the Asia Pacific] region, the software development parks in Shanghai have embedded systems context charts on the walls, and these educational and development sites are also engaging in discussions about the need for EA’s training. They wouldn’t be doing both of these things if there were not an opportunity in the marketplace for jobs.
Capgemini Program Manager Mike Ennis believes obtaining IT certifications can mean more than just having another piece of paper to hang on the wall. “I think within a few years having a testing certification is going to be just as important to a tester’s career as a project manager’s career having a PMP certification,” Ennis said.
Rebecca Staton-Reinstein, president of Advantage Leadership, concurs with Ennis. Recently, she worked with a company with clients who required a certain level of organizational maturity. “This is more and more a thing that companies are requiring of their vendors,” she said.
Sometimes, the benefits of a certification aren’t so black and white. Reinsten said doing something that isn’t a requirement shows a person is willing to go the extra mile and invest in him or herself. “There is this perception that you are more of a go getter, you are more of a self-starter,” she said. “It’s a subtle factor that I’ve found that over the years, it’s probably one of the real values.”
It’s not just an employer’s perspective that should be taken into consideration, but the test-taker’s as well. Ennis said one can learn about software testing from a best practices perspective, not just how it’s done at his or her respective organization. “It also gives you some better techniques that you can also apply to your everyday tests,” he noted.
Just like a college degree, it’s possible a professional certification can become merely another line on a person’s resume. It can be argued that it’s not about how many degrees or certifications a person has, but the knowledge he or she has and how it’s applied that matters most.
Do you think earning professional certifications is a good idea or a waste of time and money?
Organizations have heard it time and time again; application modernization can yield significant cost-savings. Just because business leaders have this fact drilled into their head doesn’t mean they know how to go about the process.
A cost-saving legacy migration and modernization project entails more than just putting a Band-Aid on old applications. Some experts say a pre-modernization assessment is a critical early step that can’t be overlooked.
Karen Tegan Padir, Progress Software’s CTO, says organizations want to conduct the assessment and create an action plan. “A modernization assessment tells you where the bottlenecks, traps, things most problematic when you are trying to move to the cloud or trying to move to a software as a service application,” she said.
Ovum IT Associate Analyst Margaret Goldberg, who researches application modernization trends, concurs. She said it’s important for organizations to take a step back and review what they’ve done and where they are looking to go. By conducting a thorough assessment, she said organizations will have a better idea of what components should be preserved and which aren’t necessary.
The careful planning can also make it easier for organizations to meet future needs. “The modernization doesn’t necessarily mean you are just going to run in the cloud now,” Tegan Padir said. “It’s the ability to have a modern app that can run and be deployed wherever you want to deploy it whether that be in the cloud or on-premises.”
Did your organization conduct a pre-modernization assessment before engaging in application modernization? If so, what did you learn?
API management company Apigee announced Wednesday it acquired InsightsOne, which specializes in predictive analytics. The move was made in anticipation of offering developers enhanced ways to “build killer apps,” according to Anant Jhingran, vice president of products at Apigee.
“The world where analytics and data were thought of as separate functions, I think is rapidly coming to an end,” Jhingran said. “The developer himself or herself has to be somebody who understands and makes use of data and insights and therefore makes his or her apps and APIs smarter.”
By utilizing predictive analytics, efforts can be more easily tracked and further refined. “If you are building an API, you really need to understand how the API is doing, who is using it, or who is not, and how you can further its adoption,” Jhingran said.
From a business perspective, the trend of taking advantage of this technology is expected to grow. In fact, research from Gartner foresees that 70% of the most high-performing companies will be using predictive analytics in the next two years.
For those who already use Apigee to build applications and APIs, they will immediately have access to InsightsOne. Jhingran said developers can now embed predictive capabilities into their APIs and applications.
The immediate benefits will also be felt by current InsightsOne customers, who will not have access to the interactions taking place through applications and APIs run by Apigee.
This has been a year of innovation and change for IT professionals, thanks to ongoing development in the cloud and mobile spheres. Mobile Backend as a Service, for example, made waves as the technology provides a fast way for developers to create and deploy the increasingly important applications.
While 2013 has been a notable year, particularly for mobile application development, what awaits us next year may prove to be even more monumental. A number of innovations will create several opportunities for IT professionals in 2014.
“This  is the year Web standards will truly go mainstream via mobile, cloud, SaaS and API technologies in areas that are beyond the retail vertical,” predicts Maneesh Joshi, SnapLogic’s director of product marketing. “The likes of Amazon Redshift and Tableau are making analytics and visualization a lot cheaper and faster to get up and running.”
APIs will become more mainstream due to connected devices and the consumerization of IT, predicts Mashery’s Delyn Simons. This is something Joshi also believes organizations need to keep in mind moving forward. “APIs are rapidly becoming the gateways to companies and information,” he said.
Experts anticipate mobile technology will continue to be a pioneering technology. A survey conducted by Forrester Research, for example, foresees mobile applications as being the most revolutionary technology in the upcoming year.
Progress’ CTO, Karen Tegan Padir, echoes the significance of mobility, noting the major impact wireless devices has. “Long gone are the days where people who work for businesses walk into a meeting with a piece of paper and a pencil and leave and wait until they get back to the office to do data entry,” she said.
Going in hand with mobile technology, it may not come as a surprise that the cloud is also expected to be a force in the New Year. “You can’t have all of these mobile devices without a nice infrastructure that is serving them,” said Padir. “Your device is only as good as how much memory you have on it. You want to be able to have the cloud services that power or fuel those.”
Do you agree with these predictions? What do you anticipate will happen technology-wise in 2014?
It’s a no-brainer that application modernization relies heavily on technical expertise, however, IT professionals may overlook the impact politics plays when determining how to best deal with legacy applications, according to Ovum associate analyst Margaret Goldberg.
Applications and a modernization projects don’t just affect an architecture team, but an organization as a whole. That means there are many people who can be for, or against, making any sort of modification to legacy applications. This can be a problem, as Goldberg noted that everyone within a business needs to be on the same page to get a project off the ground.
“Some people subscribe to this mentality if something isn’t broken, don’t’ fix it,” Goldberg said. “This mentality, along with the overall lack of growth in an IT budget, means some stakeholders won’t see the business value in these types of projects.”
Without a clear vision of the payoff an application modernization project, business leaders can be resistant to any related proposal. Additionally, some people get very attached to certain applications and simply don’t want the change to occur.
“There are some really passionate users who hold onto ‘zombie apps’ who don’t realize the resources and investments needed to service these applications might exceed value,” Goldberg said. This is where it’s important to be able to show business leaders how updating legacy applications is important to the organization’s financial health.
Getting modernization resistant individuals, especially if they are in a leadership role, to see the value of updating legacy applications is important, Goldberg said, because “you need to implement change management.”
Even if an application modernization project goes off error-free, Goldberg noted that means little if people are resistant to using the new tool.
Have you struggled with the politics of application modernization? How did you handle it?
Social media isn’t just being viewed as a way to connect with others for fun, but as an integral part of an enterprise’s makeup, research indicates. Data from IT consulting firm Tata Consultancy Services shows that since 2010, 64% of enterprises have assigned at least one employee to man social media efforts. Given this trend, enterprise architecture needs to be able to support such business initiatives.
The adoption of social media and social media applications undoubtedly affects enterprise architecture. Matt Brasier, head of consulting at C2B2 Consulting Ltd., said social media can have a big impact on SOA architecture, workload in particular. “The workloads that these kind of applications deal with is quite heavy compared to more traditional mechanisms of interacting, communicating, even sort of going back to sort of Web forums and wikis,” he said.
In addition to coping with different workload sizes, architects have to keep in mind how social media is used. For an enterprise, the scope of potential social media problems encompasses not only integration, but employee use as well. There are numerous examples of people sharing personal or proprietary information on Facebook or Twitter, only to regret such actions.
The consequences of unsanctioned data being posted on a social media platform, where it can be easily viewed and shared, can be disastrous. Another downside to social media in the workforce is the additional layer of “office politics” it adds, said Michael Orgrinz, principal architect for global markets at Bank of America.
Experts have recommendations for reducing risks and applying policies to social media tools. Ogrinz said one way to appropriately integrate social media channels is to control the flow of information. “One of the challenges is keeping these business rules current along with maintaining up-to-date information on employees’ roles and responsibilities to feed into these check-valves,” he said. Of course, such control needs to be applied delicately or there is risk of employees becoming frustrated by management.
On the other hand, social media can be a helpful tool. “One of the positive aspects though is that social media can increase an employee’s perspective about their role,” Ogrinz said. “If they are in a position that lacks direct customer interaction, social media may fill the gap and give them a better understanding of how their work impacts the overall business.”