The Question Without an Answer

What exactly is “Cloud Computing?” The term will probably never be fully fleshed out in terms of a common definition, and at the end of the day that’s fine with me. Yes, we need to put structure around the term and the industry, but as we noted earlier in the year with a post about cloud maturity models, we run the risk of painting ourselves into a corner if we try too hard to make things fit neatly into buckets that we can easily classify and categorize. Taxonomy will be key to understanding and adopting cloud computing, but I’ve become convinced that in order to truly embrace the cloud, we must…

Embrace the Abstract

I had the opportunity to speak at Interop Las Vegas this year with Rick Nucci of Boomi and R. “Ray” Wang of Forrester Research, and when I made the statement that “cloud computing is the technical manifestation of Service Oriented Architecture,” I realized that I had found the unifying principle of cloud-based solution design; unifying both in the sense that the promise of SOA finally has the technology behind it to transform it from philosophy to practical design pattern, as well as in the sense that the term “cloud computing” itself was being co-opted in much the same way that traditional software vendors co-opted the entire concept of Service Oriented Architecture to sell middleware throughout the decade.

Cloud computing may mean many things to many people, but in the end its full potential can only be realized if we stop trying to think of it in terms of the “known known” and embrace the the “known unknown.” More importantly, when we think about the cloud and applying SOA design principles, we cannot continuously innovate and drive value if we are traversing connections inward to rationalize patterns and explain the abstract with the known; we must restructure our patterns and embrace the abstract in an attempt to forge new connections by moving outward beyond our comfort zones. The next generation of system design is less about creating code, and more about assembling services – innovation through extending value in what already exists rather than inventing new sources of value. In terms of practical application, this means moving up the stack and…

Learning to Describe Rather Than Prescribe

An interesting pattern that I observed throughout 2009 is the continuing tendency to try and reinvent the wheel despite the fact that not only has the wheel already been invented, but it’s right in front of our eyes and meets our needs 99 times out of 100. I saw this over and over with Salesforce CRM and Force.com projects; the value of using Platform-as-a-Service is that someone else (or more accurately, thousands of other people) has already thought about just about everything you could possibly need in a data model, user interface, and business rules. And not only have they published just about everything you could possibly want in an easily configurable platform, they host it and manage all of the operational details such as backups, upgrades, and security. Yet time and again I encountered teams that thought that their way of doing things was better, and would go down the path of trying to build Salesforce on Salesforce before realizing that the same outcomes could have been achieved by spending a little more time upfront optimizing business processes and making minor configuration changes than going down the path of creating complex custom workflows, classes, and user interfaces to achieve the same end.

Some case studies are extreme, such as the support manager that duplicated Contacts across multiple accounts and assigned multiple portal logins to customers – in one extreme case 101 times – rather than setting up sharing rules properly; I don’t have to tell you what a data quality nightmare that ended up being. In other cases, it’s simply a matter of building rather than reusing what’s already there, resulting in hard-coding of attributes and logic that should be dynamic and extensible.

What I’ve come to realize is that there is a conceptual barrier that we need to overcome when it comes to metadata and other abstracted entities; because multitenancy architecture and SOA are reaching such a pervasive state, we must shift our thinking to describe what already exists rather than trying to recreate it. Entities exist once and in perpetuity – for example, there is only one of you in the entire world and you cannot be recreated on demand – thus our ability to provide context necessarily requires us to describe the entity in a manner that provides value to the application; the ability to describe entities with deep domain knowledge and create relationships to other entities that enrich the value of the data set will become an important competitive differentiator.

This will take time and a great deal of trial-and-error until we get it right, but in the end it is the only way to leverage the tremendous potential of core cloud computing architecture patterns; our entire concept of producing and consuming services has to change, which leads me to the conclusion that…

SaaS is Dead…Long Live SaaS

Granted, the title of this blog predicates from the acronym for Software-as-a-Service, but as cloud services mature and the traditional technology stack gets blown up and reassembled, the entire concept of “software” shifts from the self-contained, monolithic packaged application to that of a delivery framework. Software was invented to make hardware useful; hardware is abstracted in the cloud and we no longer write code down to the kernel level – we assemble, configure, and code to the layer of abstraction of the specific cloud platform. The term “software” will gradually fade from our lexicon.

This was the philosophy that drove the Java language and multi-platform virtual machine concept, and will continue to evolve with next generation rich Internet application frameworks such as Adobe Flex / AIR. What we will see moving forward is the continuing shift from desktop software that interacts with the cloud, to ubiquitous frameworks that consume data and logic services from the cloud and leverage the processing capacity of the local machine to enhance the user experience.

2009 and its economic and sociopolitical malaise are now behind us, and by all accounts 2010 will be the year of the cloud. While the technology and the terminology of the cloud have permeated the mainstream, it will take significant shifts in thought processes and design patterns before the cloud can be fully leveraged. Here’s to a great New Year and the hopes that the likes of Microsoft and other relics will accelerate their fade into obscurity and stop trying to steer the cloud discussion back into a box. Until next time, here are some…

Other 2009 Wrap-ups and 2010 Predictions

Jeff Kaplan: Key Challenges Facing Cloud Computing in 2010 and Beyond
Phil Wainewright: Tips from 2009 for a prosperous 2010
Dave Barry’s year in review: 2009 (Humorous, non-cloud related)

When I first read the OpSource Cloud announcement, I can’t say that the concept moved me to want to shout it from the rooftops; part of the reason is that I spent the first 10 years of my career managing IT infrastructure and I lost my passion for it long ago. Another reason is that my company, Delivered Innovation, lives in a different part of the cloud “stack” and we leave the management of platforms and databases to our partners such as salesforce.com. But I’ve followed Treb for the past few years, and I knew that someone with his vision had a grander plan than just building a better Amazon EC2.

While OpSource Cloud will of course be widely compared to Amazon and RackSpace at a technical level, what I was looking for from Treb during his demo was whether he attacked the major concerns, real and otherwise, that enterprises have about moving IT resources beyond the walls of their corporate data centers. The web-based management and provisioning interface had a clean look and feel, the process of setting up an account and system users was straightforward enough, and getting up and running with a standard Red Hat image took all of a few clicks and about 8 minutes of processing on the back-end. I was extremely impressed, although admittedly I don’t have enough experience with other services to provide any type of competitive benchmark. But what I did come away with was a sense of Treb’s laser-like focus on the key barriers to enterprise adoption that have caused cloud computing services to be regarded more as curious novelties than real drivers of value in the corporate IT world: security, performance, and management control. And it was this focus that got me excited about OpSource Cloud, because I see this service offering as being a key step forward in driving the legitimacy and acceptance of cloud computing in the enterprise.

My key takeaways from the demo and my discussion with Treb:

1. OpSource isn’t reinventing the wheel here – rather than using a proprietary virtualization technology, they are leveraging VMWare, which already has a significant foothold in the enterprise. This has its pros – namely that IT organizations are already familiar with VMWare and OpSource Cloud gains instant credibility by leveraging the VMWare brand equity. It also has its cons – Oracle, for example, has chosen not to embrace VMWare and does not openly support its products when installed on the platform.
2. The user security appears to provide granularity and control far beyond what Amazon provides. While I don’t think many IT folks would run off and “spend all the money on XBoxes and flat screen TV’s” as Treb joked about, the enterprise IT market demands the same type of access control that it has for its in-house platforms.
3. When I pressed Treb about my concern that despite the 100% uptime SLA and the guaranteed latency between systems, we were still talking about the same public Internet and OpSource would be in the position of having to defend itself from user dissatisfaction over any performance degredation regardless of the location or source of the issue, he laid out a very compelling vision of leveraging MPLS between the corporate site and the OpSource Cloud to provide quality of service over the public network. I’ll leave this up to folks smarter than myself to determine the feasibility of this, but I like that there was a solution in mind even though the problem may not necessarily be within the sphere of OpSource’s control.
4. Customers have complete control of private VLAN’s and network address translation (NAT), allowing them to determine which servers can be accessed over the public Internet, and which ones stay completely private.
5. A complementary storage offering is in the works, and is expected to be released in about two months.
6. OpSource Cloud will provide burstable CPU, allowing enterprises to scale computing capacity on demand.
7. Customers will be able to leverage dedicated firewall options.
8. The service touts SAS-70 compliance.
9. VMWare images will be portable between customer on-premise systems and OpSource Cloud instances.

It will be interesting to see how the market reacts to the OpSource Cloud. It’s a step in the right direction from the perspective of addressing core concerns that have dogged enterprise adoption of cloud computing services. We’ll definitely keep an eye on it and the press reaction over the coming days and months.

With regards to the testing aspect of Force.com application development and delivery, I have observed some emerging patterns that warrant further study and discussion:

1. Unit testing is real-time and automated. The Force.com IDE for the Eclipse SDK has been a solid development tool for some time, but the real-time feedback that you get using version 16.0 (Summer ‘09) of the IDE with the Eclipse Ganymede release (3.4.x) is of tremendous value. Granted, the Salesforce API will let you know if there is a problem with your code, but now we’re seeing potential issues called out in Eclipse prior to pushing code to Salesforce. If you think of unit testing in the traditional sense, where you are testing the validity of snippets of code, the very act of the Salesforce API accepting code that you push to the environment provides immediate validation in an intelligent and automated fashion. And with the line-item feedback provided by Eclipse, you can catch code validation issues long before saving to the server, significantly reducing the time required to track down and fix invalid code.
2. Integration testing has combined with QA to become a metafunction. I will be the first to admit that I initially viewed the Apex 75% code coverage requirement imposed by salesforce.com as more of an annoyance than anything. Now I have come to see test classes not as a necessary evil in deploying apps to production environments, but as a tremendous opportunity to bundle elements of integration testing and end-to-end application QA into a single overarching process. While most developers dread having to write Apex test classes, we include them in our design specifications as a best practice. If you think about it, what we are doing is taking the use cases that are fleshed out during the design process and building out test cases from the onset of the development effort; this is a fundamental shift from the traditional role of QA in outdated waterfall-based methodologies that build and execute test cases and QA plans only after development, unit testing, and integration testing are completed. I have worked with far too many organizations where QA was considered an albatross – the development effort was complete for all intents and purposes, the finish line was in sight, and here comes QA adding weeks, if not months, to the Gantt chart right before that final milestone. By making QA an iterative process that is tightly integrated with development processes, there are no downside surprises at the end of the development cycle; QA is already done along with integration testing, and it can be argued that the most important aspect of this is that QA is no longer a discreet process – it becomes a living, breathing overlay of the entire development methodology, producing test classes that live alongside the production Apex classes.
3. User Acceptance Testing is a combat sport. What we have found with many Delivered Innovation clients is that if a prototype meets functional requirements, it’s considered “good enough” at that point. The reality of the situation, 9 times out of 10, is that “good enough” is not quite good enough for a number of reasons. Namely, demonstrating a prototype in a controlled environment with use cases that are designed to show successful outcomes does not flesh out the true functional capabilities of the app; only when users start beating on an app do we start to see where the stress points are in the functioning of the application. This is why we encourage what we half-jokingly refer to as “break testing” – we encourage users to “break” the app by any reasonable means. Realistically, we want this to happen upfront and not two months after production deployment. Another reason why UAT becomes critically important in the testing process is because the execution of test classes is constrained by governors and limits far more conservative than production governors and limits. For example, while you can retrieve 10,000 records in a production (non-trigger) SOQL query, you are limited to 500 records when executing tests. While you may be able to write test classes that execute within governors and limits, once you release the app into the wild you just never know when Salesforce users will find the “unknown unkowns” and use the app in a way that could not have been anticipated…followed by the dreaded ‘Script Exception’ email alert showing up in your Inbox.

How has Force.com, or PaaS in general, impacted your testing and QA processes?