New-age Transactional Systems - Not Your Grandpa's OLTP
John Hugg discusses high volume transaction processing applications with high and low frequency profiles, and how VoltDB can be used for that purpose.
The content has been bookmarked!
There was an error bookmarking this content! Please retry.
Posted by Mark Figley on Aug 14, 2007
An interesting MSDN Developer Center blog entry gives us some insight into the definition of an Architect within (at least part of) the Microsoft culture by taking us through some of the history of the software architect role, including the driving forces that made architects a necessity, starting from the time when software architects were less necessary:
Not so long ago, programming platforms were more self-contained. From robust and complex environments like mainframe COBOL to light and easy-to-understand xBase languages, the most important technical decisions were already taken. For software projects, that meant both benefits and drawbacks. A benefit, because where a decision exists, no time needs to be spent in discussions about the best way to do anything.
Dagum, the author, then talks us through the relatively new diversity in the foundational stack architectures that we used today, and how the flexibility that those new strategies provide comes at a cost: complexity that needs to be effectively managed.
So, thanks to component-oriented software we could choose the platform which best fit a specific problem. Despite that, the paradigm shift of new platforms was difficult to learn for many developers. Thus, one of the first objectives of a good architecture was to gently hide the complexity of new platforms and APIs behind simpler ones, more focused in the context of the problem being solved (typically a business problem), and, most of all, easily understandable by regular programmers.
That way software architecture became a highway to develop business solutions in a timely, secure, accurate manner. Under the pavement of that highway is nothing but technological decisions, the critical ones. Decisions the architect had to take to make the developers’ tasks easier.
The difference with past platforms is that now, rather than buying general purpose software from a vendor, an organization’s members make the decisions regarding the function of the business solution to be developed.
Dagum explains that this precipitated a transition in the role and definition of architects, from application centric to more enterprise or even infrastructure centric roles, and even details the primary drivers behind the latest buzzword-driven architecture trend, SOA. In outlining how software architects were defined early on in the discipline, he walks through four major best practices:
He then explains how as the role has been formalized the roles and responsibilities have transitioned somewhat to a higher level of abstraction and have consolidated around three categories:
Finally, Dagum closes by looking forward, specifically referencing SOA, MDA, Software Factories, Software as a Service, and Web 2.0. The article provides an insightful overview of the development and formalization of the role of the architect and the driving forces behind it.
I think that reading Christopher Alexander's Keynote Address at OOPSLA 96 might provide an answer to the question "What is an Architect anyway?", at least it did for me.
I agree. Although Dagum's article has some interesting points, it is focused on Software Architecture and not so much what it means to be an Architect.
usually someone who can't code, but claims they used to.
Mark wrote:
> The article provides an insightful overview of
> the development and formalization of the role
> of the architect
I'd disagree. In fact, the authors of the article appear to be almost totally ignorant of the history of software architecture. It certainly didn't grow out of patterns and frameworks (although both are of course relevant to it). It's been studied as an academic discipline since the early 90s so practitioners were doing it well before people started to study it in academia.
They also appear to miss some of the really important aspects of software architecture including understanding the breadth of stakeholder needs and focusing on system qualities, not just functions.
I'm also quite puzzled by the idea that MDA is supposed to be based on SOA ...
The role of the architect is to produce valuable architecture specification. Architecture is best described through a set of views, each specifically designed for particular stakeholders. Each view provides the knobs its stakeholder wants to tweak, and hides architectural information that isn't relevant to that stakeholder.
The role of the architect is a sociotechnical one: To get the right information to the stakeholders, so that when they come around the same table they can have the right discussions and come to the right conclusions. It is to make sure that confusion and misinterpretation is arrested and addressed promptly and adequately. It helps if some of the more mechanical level of the architecture can be validated for consistency automatically.
It helps to have an architect who is a jack of all trades, and ideally master of a few also.
As Dagum says, the most important view at the moment in software is likely to be the SOA/REST view. It shows how services and their clients interact in a detailed way, and ensures that expectations and provisions of data are compatible with each other and allow each to meet their requirements.
Views outlined in the 4+1 model include the logical (reuquirements) view, process (services) view, development (source code) view, and physical (deployment) view. In addition we see views relating to API structure, and other views. These will continue to evolve over time.
Benjamin.
John Hugg discusses high volume transaction processing applications with high and low frequency profiles, and how VoltDB can be used for that purpose.
Kevlin Henney examines code samples to see what can be learned from them starting from the premise that one won’t write great code unless he knows how to read it.
Jason Ayers share the observations he made watching a team of developers collaborating in real time on the same code base, pushing XP, pair programming and continuous integration to their extremes.
Michael Snoyman presents Yesod, a web framework written in Haskell and containing a web server, templating, ORM, libraries (templating, gravatar, etc.).
Richard Kreuter and Kyle Banker on how to avoid classical RDBMS transactional systems by using compensation mechanisms, transactional messaging or transactional procedures.
Attila Szegedi talks about performance tuning Java and Scala programs at Twitter: how to approach GC problems, the importance of asynchronous I/O, when to use MySQL/Cassandra/Redis, and much more.
One category of risk that project teams need to ensure they address is business value failure – delivering a product that fails to provide value for the business investor.
InfoQ spoke to the authors of Software Systems Architecture on a couple of new topics, the System Context viewpoint and Agile, which have been added to the second edition.
6 comments
Watch Thread Reply