♦  IT Priorities Research

Risks of Service-Oriented Architecture Deployment

From a physical architecture perspective, SOA requires a layer of adapter software to be placed in front of applications, and for further orchestration software to be put in place to manage the sequencing of activities. Adding extra layers of software is never going to make an application execute more quickly! Despite this, proper analysis of the processes involved can result in an improved end-to-end experience by executing applications in parallel instead of sequentially.

The biggest performance challenge is related to the processing of messages. Messages need to be parsed for correct XML, possibly transformed to a different structure, encrypted, decrypted, and routed to the correct endpoint, with associated auditing, user identification, and message persistence tasks. This can create a very significant load on the servers, as well as increasing the volume of network traffic.

Software products are continuing to evolve to minimise this message management overhead, but a parallel trend is to move some of this load away from the servers and into special-purpose network appliances. Much as firewalls used to be implemented as server-based software, but are now invariably implemented as network appliances, there is likely to be a gradual shift of some parts of the SOA infrastructure down into the network.

from SOA Platforms: Software Infrastructure Requirements for Successful SOA Deployments (BG-0041)

Virtualization

• By YE 2010, at least 30 percent of non-desktop IT infrastructure will be virtualized (up from less than 5 percent at YE2007), driven predominantly by Server Virtualization and Application Virtualization.
• Through 2010, Server Virtualization will have the single largest impact on budgets for IT hardware and support. The second largest impact will be network virtualization.
• Through 2010, three vendors - Cisco, VMware, XenSource (now Citrix), will dominate IT Virtualization, achieving 60 percent of all new virtualization deployments.

from The Many Faces of Virtualization: Understanding a New IT Reality (ST-1510)

Keys to SOA Success

If you want services to be reused across your organisation, they have to be managed on an organisation-wide basis. If you don’t control the deployment of web services within your organisation, the result will be dozens of overlapping services that have confusingly similar names. One of our clients discovered that it had 24 services that were all called ‘Account’. Several of them were more or less identical (and therefore redundant and unnecessary) but, even worse, many of them were completely different. This confusion isn’t just an irritant, it can be very dangerous. For example, it’s very important (for example, in banking) that you know whether ‘balance’ means the average balance over the last month, the balance as at the close of business last night, or the balance as it stands at this minute.

from Architecture and Method: The Keys to SOA (OM-3503)

SaaS

Decision makers must ignore the hype and avoid the easily overlooked traps that are intrinsic to current SaaS offerings. Implementing SaaS will not be an inordinate burden on IT if management fully addresses key issues before full deployment. After selecting the best-aligned vendor and before full deployment, IT managers in larger IT shops should undertake a pilot project of part of their implementation process and spend 30 to 60 days executing the following:

• Creating an SLA management team and putting the team in place.
• Developing growth scenarios based on the field performance of the SaaS solution. This will give a truer picture of how the solution will affect the corporate bottom line than the growth numbers developed during the selection process.
• Developing and deploying data integration processes and tools that will facilitate integration with the SaaS solution.

from Pilot SaaS to Avoid an Implementation Mayday (IN-6454)

Implementing SOA

If Wave I is about learning through doing, Wave II is about sharing. Today most enterprises engaged in SOA find themselves in this transition. Having committed to SOA as an approach, these cross-departmental initiatives target business processes that necessitate collaboration among user domains and their IT groups.

Sharing of services across user domains means reaching agreement among oftenconflicting priorities; this requires the beginnings of a governance process for resolution. Consistent practice, too, in implementing cross-departmental process workflow precipitates the need for coordinating expertise, solution tactics and the reuse of services.

If vendors tout improved business flexibility as a user benefit, and users expect reduced development costs through reuse, process efficiencies and improved system maintainability, we have found these user expectations are at best, based on our SOA interviews, premature.

A more successful approach would tune user expectations to the specific nature and challenges facing them. When successful, Wave II efforts are more likely to realize software and process consistency across departments and some tactical sharing and reuse of components and services.

from SOA Reality Check: Three Waves of Adoption Through 2012 (ST-1506)

Software-as-a-Service and On-Demand Software

This isn’t really a technology in the same way as the others included on this list, but it deserves to be mentioned here nonetheless. The success of Salesforce. com has made everyone aware of the potential for service-based delivery of software, and virtually everyone in the IT space is talking about it. Results from current Info-Tech research show that two-thirds of buyers who favor on-demand software delivery believe that Software-as-a-Service (SaaS) is at least "sometimes" appropriate. Responses for desktop productivity software, content/data management software, and other applications show that about 50% believe that on-demand delivery is "sometimes", "often", or "always" appropriate.

This burgeoning acceptance of the SaaS model, and the business success of Salesforce.com, has built tremendous supplier interest in using on-demand models to reach new buyers. The combination of demand-side pull and broad supply-side push means that SaaS will be much more prominent in 2007 than it has ever been in the past.

from Seven Hot Technology Trends for 2007 (IN-6430)