elemental links

Curious about SoaML (UML Profile for SOA)?

At the March SOA Consortium meeting, Cory Casanave, CEO, Model Driven Solutions & ModelDriven.org, gave an overview presentation & demo on Enterprise SOA Modeling with the new OMG SoaML UML profile. (pdf)

SoaML is a UML profile and metamodel for the design of services within a service-oriented architecture.  SoaML can be used for architecture level modeling, or as part of a model driven architecture (MDA) process, starting with a business model and transitioning through logical and physical models, resulting in technology implementation. Since it is a UML profile, it is immediately compatible with existing UML tools.

Casanave began by setting context, describing the rationale and objectives of SoaML,  how SoaML views a service (agreement between parties to exchange something),  the  top-down (business-driven) and bottoms-up (legacy-aware) usage paths, and the mapping of those paths to model driven architecture (MDA).

To bring the specification to life, Casanave walked through the artifacts related to a claims processing scenario, including the services architecture model (see below), business process model, service contract, participant interaction model, message types, service interface, service usage, participant model, composite application structure and information model. 

[Click on Picture to Enlarge]

Included in the example was the iteration of model detail as the process moved from business concept to logical model to systems model.  In a follow-on demonstration, Casanave highlighted the transition from physical model to technology implementation using ModelPro, a new open source tool.

Even if you don’t have the time to listen to the podcast of Cory’s talk, I highly recommend flipping through the slide deck to see SoaML in the context of a real business example.  All of the artifact slides are nicely annotated like the one above.

To listen to the audio recording of the presentation portion of Casanave’s session and/or view the rest of the slides go here.  To view the entire SoaML specification, please go here. (200 page pdf)

 

[Disclosure: The SOA Consortium is a client of my company, Elemental Links.]

Business Outcomes from Services & Assemblies? Bragging rights on your S-O-A success

Did you hear SOA died again last week?  Does this make SOA more dead?   Finally dead?  Or, on par with Jason from Friday the 13th, deadish, but always lurking for a sequel?  Bad analogies aside, this time Anne is much more deliberate in separating “SOA” the marketing term, from s-o-a the practice:

"SOA" as a term has lost its luster, but "SOA" as a practice is essential for all organizations going forward” 

So, you could say I was right in my “cheeseburgers are dead, but demand for burgers with cheese is at an all time high” quip from January. 

This all brings me back to something I was talking and tweeting about last fall, that “it’s time S-O-A stood for services, outcomes and assemblies”.  As I shared with Dave Linthicum, I picked 2009 as the year organizations shift their SOA perspective from infrastructure to solutions.  And, that we’d start to hear more frequently from organizations that were early with this business-driven mindset.

This (finally) brings me to the point of this post… If you have a story to tell, now is the time.  The SOA Consortium and CIO magazine are once again partnering on a SOA case study contest.  Entries are open until June 26, 2009.  For more information on the contest, I’m excerpting my post from SOA Consortium Insights:

“Do you have a SOA Success Story?  Can you tie your SOA initiative to quantifiable business value and/or IT efficiency?  Have you learned lessons that others should know?  Did you find a way to persevere despite a tough economic climate and industry noise on SOA’s health?  Would you like an opportunity to share your story?  Garner industry-wide praise for your well-deserving team?  Then, look no further…

The SOA Consortium and CIO Magazine are proud to announce the 2009 edition of our Service Oriented Architecture (SOA) Case Study Competition. The competition is open to organizations of all sizes, including government agencies, which have successfully delivered business or mission value using a SOA approach.

Similar to the inaugural contest in 2008, the goal of the SOA Case Study Competition is to highlight business success stories and lessons learned to provide proof points and insights for other organizations considering or pursuing SOA adoption. To qualify for the competition, the SOA project must be complete with demonstrated business results.

Entries will be judged on the complexity of the business problem addressed, the ROI/Business Value achieved (Agility/Innovation/Flexibility/Optimization/Resilience), the level and sophistication of the cross-organizational collaboration (Business/Technical), the usage of SOA approaches and supporting technology and lessons learned. In addition to one overall winner, organizations will be recognized by industry/government.

To learn more about the contest and start your application, please visit the case study contest center on the SOA Consortium Site.

To read about last year’s winners go here.  And, for insights into the 2008 judging criteria, go here. 

To help us spread the word on real-world SOA success, please tell your colleagues and SOA community friends about our contest with CIO Magazine.

Fine print: Entries are open now through June 26, 2009.  Only entries from team members are valid.  [Translation: We will screen out vendor submitted marketing case studies.]”

 

[Disclosure: The SOA Consortium is a client of my company, Elemental Links.]

@ Interop’s Enterprise Cloud Summit: Fireside Chat w/Microsoft’s Amitabh Srivastava

Alistair Croll is interviewing Amitabh Srivastava, Senior Vice President, Windows Azure, Microsoft.  Amitabh is responsible for Azure, his background is core O/S.  Once again, I’ve captured some, but not all of the conversation.

AC: Are you faced now with open cloud/closed cloud dilemma?

AS: Big advantage in cloud is centralized control.  Cloud provider picks the hardware, don’t have to worry about accommodating old equipment and software.  This control and homogeneity drives down the cost.  However, still want to build a very general purpose platform.  Microsoft views cloud as extension of enterprise.  Developer can choose how to write the apps for the 3 screens – pc, phone, tv – in the environment that makes most sense.  We are still 5 minutes into the first quarter of cloud. 

…Microsoft has been doing services for years.  Have over 200 services running today.

AS:  Look at cloud as massive distributed computer scaled across globe.  Manage resources efficiently to drive Capex down.  Deliver efficient solutions to drive Opex down.  Azure gives customers the choice, on-premise or cloud, for any portion of an application.

AC: 2.4 million Microsoft Service Professionals, is Microsoft’s strategy “use the same APIs you’ve always used”? 

AS: On Azure you can use Eclipse, Python, Ruby etc.  You should be allowed to use the environment and languages you use now.  The key goal to preserve is that the skill of a developer who wants to move to the cloud should transfer.

AC: Can developers still do joins?

AS: SDS (SQL data services) allows developers to do this. 

AC: Does it scale (yes).  Why can’t Google do this?

AS: SQL data services work on partitioned database.

AC: Will existing SQL code just work?

AS: Can’t state this unequivocally.  There are certain good principles you need to follow to write scalable, available application for the cloud.  [software architecture anyone?] 

If you don’t want to make any changes, run existing SQL in VM in cloud.  If you want to achieve scale, you need to work on the application.

AC: Are ISVs rewriting for cloud?  Or, are they just running existing app in VM?

AS: Microsoft is not forcing re-write.  ISVs can port applications, or can do extra work.  Providing choice.

AC: what if I rewrite my application, and then want to move application back in-house?  Can I get an Azure cloud in-house (private cloud)?  Or, do I have to rewrite my app again?

AS: Plan is to take learning from cloud and push those into on premise offerings.  Emphasizes “over time”.  Then, the on-premise and cloud will interconnect via management services.

AC: So, your whole mission is to completely blur the line?

AS: Yes.

AC: Microsoft is late to the game, not the 800lb guerrilla..

AS: Took our time.  Wanted to consider what customers really need. 

…Big customer issues/needs are compliance, latency, trust (data mostly

…three variables – where is it, who manages it, whose policies?  answers to these questions, maps to public cloud, hybrid cloud, private cloud models

AC: Buy O/S patches are free.  Cloud, subscription model, pay as go.  How does this impact business model and partner models?

AS: Assessing this now.  Has to make sense for existing customers, existing systems integrators.

… Will be a big cultural change.  Hugely influenced by Ray Ozzie and Ray’s service vision.  This is a powerful vision, resonates with customers, gives power of choice, exposes platform for new generation of applications. 

AC: Twitter fail whale is new blue screen of death.  What happens when cloud fails?

AS: Massively distributed computing environment.  Hard computing problem.  Microsoft, Amazon and Google are all working on automating this environment.  Everyone is learning and refining systems.  Will get there.  Enterprises want SLAs with teeth.  Want recourse beyond refund.  This will evolve as cloud matures.  No one is committing to this level of service today, or even in near future.

AC: Will there be a free level to Azure, similar to Google AppEngine?  Google AppEngine free level is traffic based, not time based.

AS: This is being worked on.  Not part of Amitabh’s responsibility.  His responsibility is the technical problem, making sure the data center stays up.

AC: Organizations need to think in respect to containers rather than servers…

AS: Yes…

AC: Will Microsoft be making cloud based services of existing products, such as the analytics suite?

AS: Yes, many services offered today.  …  Fourth layer of services are “finished services”, such as Sharepoint. 

AC: Will the use of finished services breed lock-in?

AS: The underlaying platform is designed to be opened.  The higher you move up the stack, the greater potential for lock-in.  [This is no different than any on-premise business application use.  Think SAP or Oracle applications]

Audience question on cloud interoperability (inter-cloud).  Both Alistair and Amitabh are in the “it’s too early to try to standardize camp”.  Standardization (at this point) could slow down innovation.  Or, could make standards excessive – outpace actual customer needs [Think WS-*].

Audience question: Does Microsoft plan on releasing metrics on cloud offerings, database performance, etc?

AS: The beauty of cloud is anyone can write app to surface these metrics.  But yes, good question, and we do have metrics.

Audience question: Is part of business plan to let a third party run Azure?

AS: Azure will be available via server products.  Azure will not be packaged and offered as an operating system.

@ Interop’s Enterprise Cloud Summit: Fireside Chat w/Lew Moorman, Rackspace | Mosso

Alistair Croll is “interviewing” Lew Moorman, CSO, of Rackspace.  I’ve captured some, but not all, of the conversation.

AC: From your recent filing, you have 62,078 servers.  69% are cloud. Rest are managed.  What’s the difference?

LM: Cloud is set of technology about pooling and automated software.  Cloud makes shared environments robust and reliable.  Previously, in a shared environment, if someone else burst, everyone else loses.

Lew describes dedicated hosting as a house, and cloud as a hotel room.

AC: You have one set of costs associated with virtualized servers and another for dedicated.  Where do you (financially) want customers to be?

LM: Don’t break out the economics.  Cloud is more capital efficient, but very competitive which impacts margins.  In dedicated environment, customers ask for specific server and configuration.  Harder to redeploy when customer leaves.

AC: Why do customers ask for specific configurations?  Is it server hugging?

LM: Might be for security, for cutting edge technology implementations, for performance, or just preference. 

AC: What’s the biggest driver for your customers?

LM: Rationalizing your computing operations.  Economic climate, customers need to save money.

AC: Will you launch a sharded service, similar to your email service?

LM: Dedicated to Cassandra project.  Excited about Drizzle.  Needs to be a next generation datastore that people use.  Needs to be standards based.  Won’t commit to service offering.  But, data scaling without huge operational overhead is a problem customers need solved. 

… Using these new data stores and techniques is a commitment.  Need to change applications.  Service providers need to make possibilities available and help educate.  The final technology choice is up to the company.

… We are not an integrator.  There will be huge market in professional services to help people scale, efficiently.  Easy to scale, not always efficient.  This is not a market for Rackspace.

AC: Seems like engineering needs to tie into user experience more, with a conversation on accounting.  “X performance adds n cost.  Should we do this?”

LM: Dealing with IT sprawl is first order problem for enterprise IT.  Scale is less common driver.

AC: Mosso still based on VMWare?

LM: No.  Didn’t meet the needs of a service provider at massive scale.  Prefer to leverage existing, but will build solutions as required – feature/function and economics.

… Using Xen on Linux.

AC: Does this make it problematic for companies to move internal VMWare image to your cloud?

LM: There will be machine image standards in the future.  There are tools to move images between clouds.  People do move between Rackspace and Amazon today.

AC: How can you position (and survive) on ‘fanatical support’ as a utility company?

LM: Cloud computing is not going to work as perfectly as power out of outlet.  Lew calls out attendee poll yesterday, no one raised hand to “will cloud computing run perfectly”.  Also ties to Jesse Robbin’s lightening round talk on Failure Happens. 

People will want service to assist in cloud setup, run and evolution.

AC: Does cloud need reference model for assessing true cost of ownership?

LM: Glad McKinsey did study, sparked a lot of conversation.  The biggest flaw was it assumed none of benefits of cloud computing, enterprises need excess capacity. 

AC: Final advice for IT?

LM: Don’t listen to sales folks, like me.  Go use it!  Go experiment.  Solve problems that exist in your company.

@ Interop’s Enterprise Cloud Summit: Cloud Costs and Billing Models

Paying for It: Cloud Costs and Billing Models

Moderator:
Allan Leinwand, Venture Partner, Panorama Capital

Speakers:

• Thorsten von Eicken, CTO and Founder, Rightscale
• Grace Kim, Sr. Manager, Marketing, WebEx (Cisco)
• Jesse Robbins, Co-Founder and CEO, Opscode
• Richard Dym, Chief Marketing Officer, OpSource

Allan Leinwand opens with telecommunications innovation story, how “friends and family” plan broke AT&T’s monopoly and points out that “friends and family” is back now in cell phones. As such, he sub-heads the panel as “Why Pricing Matters”.

What do you use as pricing model for cloud based service, has it evolved over time?

Webex has 3 pricing models: named user model, ports model, minutes model.  Ports model is attractive with global companies, port use follows the sun.

Rightscale has a per hour model, but there is a floor in place, so minimum buy-in.  The floor covers start-up expense.

Opscode hasn’t publicly released product.  Jesse is very interested in metered billing models.

OpSource hasn’t released product either, cpu hours with monthly billing is probable.  believe enterprises will look for annual type contracts with bulk purchase discounts.

Allan points out that base unit isn’t at a level the CIO typically thinks (plans).

Richard – Amazon has done phenomenal job and has set the tone for units and billing.

Thorsten – What’s important is the visibility and control, not the base unit. 

Jesse – The kind of monitoring and metrics that exist in the cloud, should exist in the enterprise.  This is a huge opportunity for anyone building or operating infrastructure should start thinking about this now, then you can do an apples to apples comparison of on-premise versus cloud.  [Jesse used to run Amazon’s operations]

Need to understand more than aggregate install, need to understand usage as percent of aggregate and remove the waste.  This is an opportunity with virtualization.

Thorsten – The visibility provided by the bill often triggers application optimization efforts. 

Length of contracts?  What type of bill should people expect?

WebEx is typically a one-year contract.  There are options to sign-up for shorter term, higher price.  WebEx has free trials and pay per use models.

Rightscale started with one-year contract and monthly billing.  Became administration issue.  Need to align contract and billing.  Rightscale does their own metering, but outsources the billing.

Jesse sees movement towards no upfront contract, ease of startup, then transition from free to pay as usage grows, at this point, the enterprise has bought into the solution, and you remove the long pre-sales contract work.

Would you ever offer flat based pricing?  Allan cites recent trend in telecommunications from variable to flat.

Jesse – depends on the service being offered.

OpSource – yes, can see this evolution. 

Audience

Q: Do you use any vendor tools for metering or billing

A: Some providers: Aria systems, Zuora, Evapt, OpSource

Q: Are SLA based refunds automatically tied to billing system?

A: Don’t believe so. 

A: Free doesn’t require refunds – Jesse

Bottom line from panelists: “pricing can be fun” or “like eating glass”.