SOA, Cloud and the Rise of the IT services-based Economy
So, are expectations for SOA and Cloud computing over hyped?
A SLA is a legally binding contract that defines the relationship between the service provider and service consumer. As an individual cell phone user, you most likely have such an agreement with your mobile service provider, but, unfortunately, it favors them, not you. However, a business cannot afford to agree to such unfavorable terms, and thus, must enter into contract negotiations to ensure that the service provider delivers a minimum level of service or is penalized.
Service providers, on the other hand, have no choice but to submit to these negotiations if they wish to acquire the business. In return for committing to delivering a minimum level of service, the service provider often will impose termination fees to cover the costs of the overhead associated with entering into agreement with the service consumer should they change their minds before the expiration date on the contract.
Thus, businesses are highly unlikely to enter into service consumer scenarios where there is no commitment to ensure availability of the service relative to demand. If demand is low and non-mission-critical, then a business may not bother at all with an SLA. If demand is high and/or mission-critical, the business will not move forward without a carefully negotiated SLA. (Of note, the cost to develop an SLA averages between $30,000 and $50,000.)
SOA and Cloud
Enter SOA and Cloud computing. SOA represents the archetype for delivery of everything as a service: software, hardware, telecommunications, call center, marketing, etc. If something can be compartmentalized with a formalized input and result, and can be measured, then it can be delivered as a service. Services working together to provide a more comprehensive offering is SOA. Cloud computing is the extension of this archetype for delivery of computing services, typically over the public Internet.
Ubiquitous availability of low-cost, high-speed Internet connections is a disruptive and driving factor toward the rise of an IT services-based economy. Prior to having widespread access to high-speed bandwidth via broadband, mobile and wireless connections, companies, typically, physically acquired software and hardware, installed and managed it locally. This put the onus of support on the business itself backed up by the vendor usually under peripheral support contracts. In some cases, where businesses had acquired capabilities as a service, it required expensive leased connections, tape-based data transfers and costly transaction fees.
With the rise of an affordable and commoditized approach to delivery of software and hardware as services available over a public, networked connection, businesses can now envision a future in which they can gain the advantages associated with an IT services-based economy―thus reducing the overhead to their own business without sacrificing capability. Yet, for this model to truly succeed it is going to have to address the needs of businesses with regard to service levels without incurring the overhead associated with SLAs. Moreover, as services become dependent on other services, the aggregation of multiple SLAs is going to further complicate this process.
While it seems plausible to create a one size fits all SLA, the realities of corporate performance, auditing and compliance with governance standards, such as Sarbanes-Oxley, are just some of the factors that impede the likelihood for the emergence of this type of agreement. Also, as noted above, Cloud computing is being developed in compounded layers, each with their own SLAs, thus restraining the ability for some Cloud vendors to even commit to some customer's SLA demands.However, even if the SLA issue was neutralized, businesses would still need to relinquish some control over data in order to gain the value of participating in the IT services-economy. This last point is currently a key impediment to businesses rushing to embrace the Cloud.
Let's return to the point made earlier regarding when an SLA may or may not be required, and let's call this the demand spectrum. Again, on the one end is low demand, which may require no SLA at all and on the other end is mission-critical, which requires expensively negotiated SLA.
Based on this spectrum, it would seem that the most likely candidates for applications deployed without an SLA will be those for which there is little to no penalization for lack of availability. What types of applications correspond to this classification? Emerging SaaS applications with few customers, social networking applications (e.g., Facebook, LinkedIn, MySpace, Twitter), marketing websites and public service websites (e.g., local, state and Federal government). Essentially, this is the same market for hosting, so the advent of SOA and Cloud computing does nothing to foster the rise of an IT services-based economy regarding these candidates.
With regard to the mission-critical end of the spectrum, in order for an IT services-based economy to move toward commoditization it needs to remove the reliance upon expensive, complex negotiated contracts. The rise of Cloud computing is fostering many more vendors offering computing resources as a service, which is driving costs down, but many of these vendors don't have the required infrastructure to ensure compliance with a mission-critical SLA.
This brings us to the big question, What do SOA and Cloud computing bring to the table with regard to fostering an IT services-based economy? The answer, for now, is vision and technology advancement. Unfortunately, in order for the IT services-based economy to reach mainstream, we will most likely require tort reform, policy changes and more precedents in legal cases to help ensure the rights of service consumers. That said, the introduction of SOA and Cloud computing today is laying the groundwork for a future commodity market.
In the interim, businesses should focus on leveraging SOA internally as a means of preparing for a possible launch into the Cloud. This effort is complex and will most likely take time to complete, allowing for the Cloud to mature. Internal IT organizations can also benefit from developing an IT services-based economy within their own businesses as it will provide the agility the businesses needs to optimize IT expenditures and leverage emerging Cloud options as they arise.
JP Morgenthal is a leading independent IT architecture consultant in the Washington, DC region who focuses on enterprise architecture, SOA, BPM and cloud computing. He is an authority on XML, Java, SOA, EAI and EII. He has written three books on integration, the most recent Enterprise Information Integration: A Pragmatic Approach.