Disruptor in the Data Center

In the last few years,I have had opportunity to work with with Enterprise Architects in Infrastructure Management Teams during the deployment of new systems/applications. The key observation was that we spent lot of time discussing the application sizing requirements, security requirements and network requirements to arrive at a suitable Solution Architecture [infrastructure] that would evenutally solve the problem at hand.

The actual 'System Integration' between the hardware resources like servers, storage, network/firewall typically needed to be hand-crafted and hence took significant time and effort. The scenario would be much more complicated if we were talking about Internet facing applications or the complexity of the application was high.

Coming from the Application/System Architecture background, I am thinking - Wouldn't be possible to come out with several design/architectural patterns in Solution Architectures [Deployment/Infrastructure architecture] as well. For example, Internet transactional system can have a pre-integrated topology and intranet system can have its own pre-defined / prescriptive topology and based on a standard model and the additional discussions can be had according to the very specific project requirements.

Am yet to talk to my Infra Architect counterparts to understand the viability of such a thought process.

But, It looks like Cisco has observed this very practical problem prevailing in the data center and come out with an 'Architectural approach' to solve this issue. Cisco's new offering 'Unified Computing System' is not centered towards a specific solution architecture. It is positioned as a holistic data center platform where many such 'Solution Architectures' can be crafted and deployed seamlessly.

In this new offering, the Platform is a hardware + software combination that has the 'inherent' capability to orchestrate server, storage and virtualization with high-speed network equipments. Yes, I said 'Inherent' capability with no 'plumbing'.

For the first time, Cisco is entering into the data center market, not as a network gear provider but the 'Server provider'. And Cisco has taken this daring move disrupting its own long-term partners such as IBM and HP.

By providing a 'Pre-integrated/Architected System', Cisco aims to eliminate the manual system integration that typically is the case in data center operations. Am personally excited with the 'Architecture approach' as it attempts to think out-of-the-box. [hardware box :-) ]

Cisco CTO claims that today's data centers are fragmented and innovations like virtualization are not enough. Applying those innovations would only increase the operating expenses, She states. I couldn't agree more!.

To support Enterprise data center requirements, Cisco has roped in several partners such as EMC and VMWare to work in a unified Environment. The UCS System will have the capability to seamlessly orchestrate various data center resources from its partners, thus resulting in a plug-n-play environment. With this approach, Cisco's UCS emerges as the 'Brain' in the Enterprise Nervous system [Infrastructure layer].

When I read the offering in detail, I draw strong similarities between Cisco's Unified Computing System and the SOA principles. The UCS effectively implements the SOA principles to seamlessly orchestrate different heterogenous resources using industry standard protocols.

This is clearly a disruptor to the data center/infrastructure industry. And Cisco actually in the path of realizing Sun Microsystem's Visionary statement - 'Network is the Computer'.

What is common between SOA and BI projects?

(This is one of my favorite topics, so please bear with the long post…)

Both kinds of projects fail to deliver their promises most of the times…! :-)

Here are some evidences – Anne from Burton group reports that SOA is not working in most of the organizations. She also adds that many companies have stunningly beautiful SOA infrastructure and deployed the best technology. Yet those initiatives fall out. I can personally vouch for this statement. I have seen the same experience in the recent past. And more importantly, she says Techies have not been able to communicate the value of SOA to business partners!. Absolutely true! Out of the interviews she conducted so far, She confirms there is only one company that can be classified to have done true SOA.

What a startling observation?

While System Integrators are boasting of 100s and even 1000s of SOA case studies, an industry analyst says there is only one. It all boils down to the question of - do we understand SOA enough?

Now coming to BI projects, my fellow architect in the team confirms 40% of BI projects typically fail to meet their objectives purely because of data quality, data ownership issues. 40% is not a small number, especially when we deploy fancy DW/BI tools in place.

Now, what is so common between these projects?

As I have written in mashup posts, anything beyond plain vanilla operations – say Operational Excellence, Generating Insights/Analytics, Growth, Innovation, Governance, Reuse and Sharing – requires not just ‘tools and technologies’ but the ‘people’s hearts and soul/Passion & Energy’.

If the SOA/BI/Mashup initiatives do not culturally transform the company in terms of org structure & collaboration, they are bound to fail immediately.

As I have written a comment in Anne’s post, these projects to take off and provide real business value, it requires change agents in the company. And IT architects/ consultants and Project Managers do not have enough credibility or skills in performing the role of change agents and that’s where the root cause of the problem lies. To perform real changes, it requires a leader who understands the potential of SOA/BI projects, articulates effectively to stakeholders and persuades the people to buy it for their goodness. And that’s not easy!. And that is typically thankless job!

Some of the similarities I could list:
1. SOA deals with owenership of services. BI has to deal with Ownership of data. When the projects need to deal with multiple owners for a single initiative where they have no immediate benefits, there is a challenge
2. Semantics is important to SOA. Data Quality and Semantics are both important to BI projects. Both cannot be achieved easily / automated easily and requires peoples’ knowledge and collaboration.
3. Both initiatives require Strong Governance processes
4. Both the initiatives have challenges in terms of defining the ROI and achieving it pragmatically.
And there could be few more…

Oracle Fusion Apps in 2010 : Anyone waiting?

Are you surprised that Oracle Fusion was not in the limelight during the recent Oracle OpenWorld 2008 event?. Well, it looks like there are valid reasons. For one, Oracle Fusion Apps may not be available to the customers until 2010. And the rest of the reasons follow...
I have always questioned the value of 'Big SOA' as called by ERP vendors. And this recent CIO article only helps to validate my opinion. Some of the important observations from the article:
1. It is going to be hard to justify a ROI for the new product (Fusion)
2. There is no competitive pressure from other vendors like Microsoft and SAP to achieve a faster time-to-market
3. Above all, there is no pressure / demand from customers for the new product
Same goes with SAP's SOA story as well. But, SAP is committing billions of dollars towards revamping their product and creating a completely new breed of ecosystem to support ESOA. But, What is the business case for an existing SAP ERP customer to move to ESOA ecosystem?. I have not found a compelling answer yet.
Given the fact that SAP/Microsoft and Oracle are delaying their SOA offerings,Guess who is benefitting from it?. Its none other than IBM. If an existing ERP customer wants to do SOA 'today', the cohesive & comprehensive solution could be built out using IBM SOA infrastructure.
Am not trying to sell IBM products here. But the message I wanted to pass across is that 'There is no SOA without a compelling business case. And Make sure you have one before deep diving into SOA, especially Big SOA'.

Visual SOA on your way!

Have you ever wondered why Google Maps leads over others in the mashups world?. Why do you think iPhone is so attractive over other phones? In India, iPhone 3G was recently launched and it costs $800, twice as much as the US price and people are still buying it. Why? - I strongly believe it is because of its visual appeal, ease of use and the purpose it serves. The user interface doesn't make one to 'think' to figure out what purpose it serves and how to use it. And thats the quality of good UI.
Now, let us come back to SOA world. Have you ever convinced a business analyst about why SOA is required? or beneficial to the company?. If you have had, you would know what I am talking about.
Whether its an Operational SLA explanation, product issue, or new strategic initiative like SOA - The communication between IT and business holds the key. If you are unable to communicate the technical message that is understandable and appreciatable to the business, your success is uncertain. Especially in the case of SOA, without business involvement, its just not qualified to be called as 'SOA'.
Take the case of next level from business owners to business process analysts. I couldnt think of business analysts modeling complicated business process diagrams using the famous, sophisticated modeling tools. I have interacted with those business analysts in SOA initiatives and their experience most of the times is found to be 'messy'. Though they may be able to create and author certain artificats for a while, ongoing maintenance and governance is a nightmare in those modeling tools.
So what is the solution?. - Zapthink has a take here. Zapthink is one of my favorite and admired analysts in the SOA space. And I have heard its lead Analyst Jason Bloomberg speak in one of the events. Zapthink is vendor agnostic and approaches SOA purely from business drivers perspective.
In this article, Zapthink analyst Jason Bloomberg ideates that applying 'Semiotics' to the SOA will ease the communication of Service abstraction to business. Semiotics is a study of sign processes using signs and symbols and typically used in communication. In this article, Jason gives a simple but an illuminating example of how we all use the 'Window' symbol in our Desktop to manage applications within the System.
The point Jason makes is that SOA will not be useful to the business without working visualizations. And we have a living example - Mashups. Consumers all around the world have created so many visual mashups only to prove that, if they find something visually appealing and understandable, they can put their imagniative ideas around it and create a all new business purpose serving mashups. Can we create this magic in the enterprise world?. Can we expose our business services in a Visual appealing environment?. Not in the form of 'Service repositories'.
If we can, and I would call that as SOA tipping point.
And one can see for himself the progressions being made in the visual environments. We are moving from web 1.0 to web 2.0/RIA. And from web 2.0 to Virtual worlds. Intel is actively working on a whole new paradigm called 'Visual computing' to provide a new visual computing environment for graphically intensive computing environments such as Gaming and Engineering.
I strongly believe there is one disruptive innovation that is waiting to happen between RIA and Virtual Worlds. Virtual Worlds interface is just processor-hungry. (When I tried Second Life in my Laptop, it simply hung). RIA is so good but I find it not so mature to front-end business applications. So an interim technology that brings the benefits RIA/Virtual World and a mature programming model coupled with SOA Semiotics is waiting to create a magic.

A new language for SOA!

While the end objective of SOA is to achieve enterprise agility, the actual path to achieving the same is 'counter-agile', as Thomas Erl states. The actual process of building and hosting the services in a distributed environment is not simple. And on top of it, if you add the products that promise to implement successful SOA, the complexity will only multiply.
This post is a continuation of my two previous posts - IT industrialization and Product Line Architecture. When I was discussing with my fellow Architect in this regard, we concluded that the need of the hour is 'simplicity'. And I can no way refer any of the existing products in this category.
And there is also lot of discussion going on in the web about the promise of SOA. Someone has asked, if SOA is evolutionary and promising to bring tremendous benefits to the enterprise, how come we are building SOA with same old technologies?. - very thought provoking. You may say - Web services is new technology. But, the fundamentals are still the same.
The only way I see the complexity can be reduced is by bringing a fundamentally new technology or language that would simplify 'building' SOA. And I am not wrong!. Check out this language - Somebody is not only thinking but working on those lines.
Another way of looking at simplicity is use of 'Domain Specific Languages'. I see DSLs are the only thing that is common between Agile Camp and SOA. Both of them agree to gain benefit out of them. While the 'Agile' camp approaches DSL for productivity/Accleration in Development cycle, SOA approaches DSL from 'semantics' perspective. Both stand to gain from it. I believe well designed DSLs would make the 'building' part of SOA lot more 'faster', not necessarily 'simpler'.
But unfortunately, it looks like DSLs are going to become obsolete even before they mature! Yes, Microsoft is planning to adopt UML instead of DSLs for modeling SOA.

SOA Vs Product Line Architecture

While I was contemplating with an idea of Product-Line Architecture concept to a series of projects in my company, it triggered me the thought of comparing the same with Service-Oriented Architecture. I plunged myself into this comparison of Product-Line Architecture and Service-Oriented Architecture and the journey was quite interesting and insightful.
I also saw that CMU is doing lot of work in bringing the synergies between the two paradigms. The university views that both paradigms can complement each other and each can add value to the other at broader architecture level - Applying Service-Orientation to the Product-Lines and Applying Product-Line approach to Service-Orientation / Composition. While I completely agree with this idea, I believe a greater synergy can be achieved when they are applied at their respective contexts, where they are best at.
To explain a bit more, in one of the scenarios, SOA and Product Line Approach can perfectly co-exist, where SOA addresses the interface part of Services and Product-Line addresses the implementation of components behind the services. The Product Line concepts can be applied to model the service inventory within a specific domain in the enterprise. Applying Product-Line concepts to the implementation portion of SOA will significantly benefit in terms of accelerating the development time of services.

Rise of new ERP

Came across this interesting article titled 'The Next Revolution in Productivity' in Harvard Business Review few days ago. The article advocates that - To harvest the maximum benefits out of an SOA initiative, it is required for the companies to revisit the fundamental design of their operations & their organizational structure. If they dont do that, all they will achieve is just IT efficiency, while SOA's true potential can achieve lot more than that.
This perspective generates a lot of thought process. I always tried to understand the value proposition behind the SOA strategy of bigger ERP vendors. The ERP vendors' SOA stack primarily unwraps its huge monolithic bundle in the form of prolific services and encourages the ecosystem to innovate and compose new business processes. However, I would see this model as 'innovation' happening on the 'Edge'. The true innovation in this space would be the one that could be applied at the 'roots' of the business processes.
As mentioned in the above article, using ERP vendors' SOA stack, is it possible to fundamentally redesign a specific business process?. Is it possible hot swap, sell or buy the inner most activities of a business process that is composed around ERP?. If the HBR hypothesis cannot be delivered out of today's ERP ecosystems, then as the article states, one may not leverage the fullest potential of SOA. And the one that supports the complete enterprise transformation with innovative business processes would be the new ERP...

Virtualization overtakes SOA?

Am not trying to draw any parallels between Virtualization Technology and SOA Architecture here. Am classifying both of them in the category of emerging trends and trying to see which one is progressing faster.
Though I am an ardent fan of SOA and the potential impact it could create on the IT and Business, I should acknowledge that there are significant complexities involved in implementing SOA. Zapthink partner David Linthicum talks about it persistently in his blog. SOA makes tall promises. However, to realize the promises, it requires skilled SOA architects/designers, business consultants, sponsors and above all patience. :-)
Having said that, I should say Virtualization leads the game in terms of realizing the promises and achieving the benefits. Here is a reference. PC Quest has recently announced the Top IT implementations for the year 2008 and I couldnt find a single SOA focused project. Almost all of them are pointed solutions that are targeted towards solving a specific business problem. Am not saying that pointed solutions are bad. The point here is SOA has not been mentioned as a key stragic enabler in any of those implementations.
However, I could see a virtualization success story. ICICI bank has recently completed the enterprise server virtualization project and this project has been selected as one of the best implementations. More than the project, the metrics made me to say Wow!. Here it is - The bank was able to condense more than 550 servers in about 43 servers using VMWare Virtualization Technology. This is a strong metric that speaks the benefits of Virtualization.
Virtualization as a technology concept is slowly making in-roads into the industry and Even conservative IT shops are bound to invest in this technology for the solid proof statements that can be seen around the industry.
There are plenty of SOA case studies as well. Am not denying that. However, I strongly believe that virtualization has a direct impact on the fundamental economics of an IT shop and hence it succeeds, whereas SOA has an indirect impact via the business solutions and the actual architecture that gets implemented.

So My suggestion would be - Invest in virtualization today to save money so that You can invest in SOA tomorrow! :-)

The value of EDA

Have not realized the value of EDA (Event-Driven Architecture) so much until recently.
I was always thinking whatz so great about EDA when your J2EE MDBs were supporting just that.
But, We are designing an EDA project right now and I could see the immense potential of the architecture. And there is no wonder people call it SOA 2.0 = SOA + EDA.
Truly, the next gen of business applications need to be designed in SOA 2.0 mode to ensure the business agility and performance.
While SOA relates to business functions, EDA necessarily models the business events and its consequences on business services.
The EDA addresses different levels of complexity right from Simple Event processing to CEP. While the traditional MDBs were designed to support simple event handling mechanisms, EDAs are designed to handle end-to-end event centric architecture and design of business solutions.

Not Doing SOA can invite Lawsuites?

Very interesting perspective!

Why would someone do SOA?...Two reasons:

1. To make your IT environment simplified
2. To fuel growth initiatives using the Agile IT enabled by SOA

Now, what if the corporates who have implemented SOA are more profitable than the corporates who have not implemented?. The farmer will be able to deliver more value to stakeholders than later...

David Linthicum of Zapthink group has this very interesting insight into SOA Vs Shareholder Value in the InfoWorld.

I couldn't agree more to his point. The Days are not far where the corporates are audited not just for their financial controls and security procedures, but also for effective IT systems & enterprise architecture. And if the audit reports that a company's IT is ineffective, will the stakeholders keep quiet?

So, its time for corporates to question themselves...Whether SOA makes sense for them?

Strategic Thinking Pays off!

On reading a recent story on 'Relationship based pricing' in Financial Services. This concept basically talks about offering special prices based on the relationship that a customer has with a Bank across various business units like Insurance, credit card, etc. Traditionally, the prices/discounts are extended to the client only based on a specific product relationship. For example, a Bank will not typically offer special prices when the customer makes a product purchase across a line-of-business. This is due to the hard fact that the Bank does not have 360 degree view of the customer. Hard hitting indeed!.

Every bank is trying to get this full view of the customer. There are lots of point-to-point integrations are implemented in a Bank's IT system to realize this sceanario. The point is that it exists. But, it is also complex and erroneous. So, the banks have so far not realized the complete potential of having 360 degree of customers. New products/services on top of a 360 degree view have always remained as dreams.

Now, I see this scenario as the real business case for having an SOA based infrastructure and integration. Had the Banks had an agile infrastructure today, they would have been in a position to offer new services like Relationship based pricing. Now, since the fact that their infrastructure is not agile, its going to increase their time-to-market. And thats where the competitor wins!.

So, the future lies with the people who are really acting TODAY.

Applying SOA to business, not IT

SOA is getting enough attention these days...by vendors and clients...

Recently one of my ex-colleagues, a fellow architect, surprised me by throwing the idea of applying SOA to a real business context and not IT systems. Sounds strange!?. Not really!

SOA principles are mainly applied to IT systems to make it more structured and be agile.

How about applying the same principles to the real business?. The business organization, business processes and services offered by the business...

My colleague was quoting the example of service-orienting the IT services organization. By and large, all IT services are grown to an extent where their sheer portfolio of services (e.g. application development and maintenance, infrastructure management) themselves have become an independent organization within the bigger organization. Each Line-Of-Business within the IT services organization has become siloed, powerful, bureaucratic and rigid. Introducing any change in the respective LOB or composing/creating a new service by mixing/matching with another LOB has become a nightmare!. Its hard to believe!. But its the truth...Its increasingly becoming difficult to orchestrate services within the one and the same organization though they all belong to the same parent. But, the irony is, all such IT services/consulting organization do enough cross-selling and up-selling and claim to be the one stop shop for all services.

So, How do we solve the problem?...We can look at what SAP consulting does...It has defined productized services for several of their service offerings by implementing standardized tools, methodologies, and frameworks and trainings for its own consultants across the globe. So, when a new business comes from the client, SAP consulting nicely distributes the work across the globe and orchestrates and gets it done. All becomes possible because its services/offerings are industrialised.

I can visualize a 'Enterprise-level Global Service Bus' that is running within such an organization, where each LOB is seamlessly integrated and presents their services...It feels good to imagine such a nice-and-beautiful system. But its indeed looks feasible!.