Category — Business Service Management

What’s the state of operationalizing Service Oriented Architecture (SOA) monitoring and management tools? Are the typical network, systems, enterprise operations/management centers (NOC/SOC/EOC/EMC) up to speed on how to manage, monitor, triage, troubleshoot and in general understand how SOA is being used in companies that are adopting it? Should the operations center care that they have an event from something related to SOA infrastructure and respond differently than they would for a non-SOA event? Have SOA events, incidents, problems, process and workflow been thoroughly implemented in such a way that “it just works” like traditional enterprise monitoring and management? Or, are these fancy SOA monitoring and management solutions really reserved for those applications experts responsible for complex application support and development?

If a client continues to struggle with fundamental e2e service monitoring and management, transaction monitoring and management or even batch job monitoring and management, what will their chances of success be for SOA monitoring and management? Could SOA and associated “service or transaction oriented monitoring” be a catalyst to shore up these other areas? Should one be tackled/improved before starting on another? At a minimum, instituting a “service oriented” organizational structure and mentality is certainly something I’d recommend for anyone adopting broad based SOA principles.

Eric Roch offers some solid advice on SOA Monitoring and Management which highlights that there’s more need for doing the fundamentals of systems, application and service management and monitoring really well as a foundation for SOA Monitoring and Management.

Others (and my preferred focus area) feel that monitoring SOA should really be more closely related to monitoring what this SOA initiative and deployment’s all about - the business. Business Service Management (BSM), Business Activity Monitoring (BAM) and Business Process Management (BPM) all play a key role in helping understand how IT infrastructure, systems, applications, etc. support and impact the business’s goals and objectives. The fairly new buzzword Business Transaction Management (BTM) spearheaded by Correlsense and OpTier really speaks to the desired need here.

I feel that it’s got to be a focus on both of these areas, but with a strong preference to the “B” buzzword set since most IT organizations are likely using the “improve/standardize/reuse/efficiency/time-to-market” spin to aide in business support and justification for their SOA initiatives. That said, you’d BETTER focus on the things that the business cares about and show them tangible evidence that your SOA initiative is making things better for them. This is very possible by adopting a BSM, BAM, BPM, BTM (or a preferred combination) strategy that focuses on providing the right level of business visibility into the SOA environment and more importantly the e2e business services, applications, transactions, processes and activities. It ultimately all ties back to the service level agreements delivered to the business anyway right?

What’s on the market these days for SOA Monitoring and Management? Should you get your monitoring and management tooling from your core SOA platform vendor or should you take a third party, “best of breed” approach? Are there true “vendor neutral” solutions out there? Are clients implementing SOA architectures based on multiple vendor’s technology, solutions and products?

• Progress/Actional
• Amberpoint
• Avicode
• CA Wily
• Correlsense
• HP
• IBM Tivoli ITCAM for SOA
• Manged Methods JaxView Suite
• MuleSource Galaxy and HQ
• Nastel Autopilot
• OpTier
• Oracle/BEA
• Progress (Actional, Sonic, Mindreef)
• SOA Software
• Symphoniq
• Tidal Software
• Truviso

Some additional content on some of these vendor solutions is available here.

Who might the “market leader” be of these SOA specific solutions? What makes them a leader? What capabilities, features, functions would be considered “best of breed”, differentiator, must have, core, desired, nice to have, etc.?

What’s “really” needed for SOA monitoring and management?

1. Web Services
2. ESB
3. Transaction Performance
4. Transaction Availability
5. Transaction State/Status
6. SOA Registry
7. SOA Security
8. Service Discovery and Relationship/Dependency Mapping
9. Transaction Discovery and Mapping

Anything else missing here? What here needs to be specialized in its own product versus just extending the investments clients have already made?

Please do share your thoughts here. There are folks lurking who really need help in figuring this stuff out and/or improving products and capabilities on the market today!

A new IF is available for TBSM v4.1.1 addressing a few new areas (don’t see mine in there!). This depends on IF 001 and supersedes IF 004, 005 and 006. IF 007 can be downloaded here.

These are the new issues addressed:

IZ15914 INCONSISTENT SERVICE NAME TRUNCATIONS IN SERVICE TREE
Service name truncation is not consistent when using the static sizing tool. Many of the service names will truncate, but some do not.

IZ19833(NGF) NEED THE TBSM LOGON SCREEN TO BE ABLE TO ACCEPT MORE THAN 16 CHARACTERS
Integrated authentication works only if the password is short and when a longer password is used it causes the account in Active Directory to be locked.

This APAR increases the allowable length of the password to 127 characters - which is the Active Directory limit.

IZ20375 CREATING CUSTOM CANVASSES USING IE CAUSES HANG WHEN USING DE
When trying to either create a new custom canvas or edit an existing custom canvas, the canvas would fail during loading and the console would show the following error:

[Fatal Error] ServInst.xml:2:64: White spaces are required between
publicId and systemId. Exception in thread “Thread-12″
java.lang.NoSuchMethodError:

IZ24515 UNABLE TO SAVE SERVICE INSTANCES
The customer edits a service and clicks on the dependencies tab when the customer adds or removes dependent services, the change cannot be saved. The save screen will show up when the save button
is pressed, however, it will never go away and changes made are not saved.

IZ26602 PERSISTENT ESDA INSTANCE HAS A PARENT RULE FAILS TO SHOW UP IN
When there are persistent ESDA instances that belong to a template with a parent rule, they will not be displayed at the root of the tree even if they have no actual parents. Thus there will be no way
to see them.

This fix only applies to persistent ESDA instances that has no parents. If you want to enable the old behavior to not show the instances in the root of the tree set the following:

In RAD_sla.props
impact.sla.showesdainstanceswithparentrules=false

The default for this property is true, which will show these instances at the root of the tree.

With all of the sudden hype about something that’s debated to have been around for quite a while, I suppose I should jot down some of my thoughts about how I feel Business Service Management (BSM) should play a prominent role in the Cloud Provider and Cloud Consumer spaces. I certainly hope that cloud computing could be a catalyst for broader based BSM adoption.

While I’m not the subject matter expert in all of the latest cloud, fog, haze or other buzzword laden technology talk in this area, I do strongly feel my views of BSM are as relevant here as they are in the traditional spaces we see BSM applied.

If you look through to cloud computing providers with the traditional IT management context, we’ll likely see numerous domain specific tools to help manage the servers, appliances, virtualization, storage, network, security, load balancers, etc. and various provisioning and automation capabilities within the cloud infrastructure of a Cloud Provider. There’s probably not much that would change here, other than a few new product entrants or fancy reuse/repurposing of existing products, protocols and what not. The tools provided are all relevant and needed to help solve domain specific problems by the various silos in the typical IT/Service Provider organization.

For the Cloud Consumer, we’d expect to see much more focus on an “outside-in” management approach that’s heavily focused on availability, performance, capacity, utilization and end user experience. We’ve seen the initial entrants here with Hyperic’s CloudStatus and Amazon’s AWS Service Health Dashboard. These are both good first starts at providing visibility into “the cloud” but … they somehow leave me wanting more. Finding the “right” balance of visibility into the cloud infrastructure will be something that the Cloud Providers will need to think about.

BSM. Think, Operate and Respond Different.

If you look into the cloud through the context discussed above, it’s difficult to fully relate things to the business of offering cloud services, and it’s especially difficult understanding how cloud services may impact Cloud Consumers and their businesses. IT management within the domain specific areas is about IT, keeping the infrastructure up and running is generally job one. This becomes even more important in a heavily virtualized environment like a cloud where you may have dozens or hundreds of “end user” businesses associated with a single cloud infrastructure component, chassis or rack in a datacenter.

When you operate with an understanding of the business, its goals and objectives, and how the cloud can impact those of the Cloud Provider and those of the Cloud Consumer, you have the opportunity to “think, operate and respond differently” than you may have previously if you didn’t have this contextual information.

Here are some of my initial thoughts here, more to come as I can dialog with others in the cloud space about what’s possible, practical, etc.

Cloud Provider

• When you build a cloud, provide some way to incorporate fundamental information into your monitoring and management tools that will enable the support, operations or engineering teams to understand the business impact to your company and those who are subscribing to your cloud services. This may be something as simple as a run book that contains key business and subscriber information that’s easily referenced automatically or manually by these teams. If you have a fancy autonomic provisioning or capacity adding wizard, ensure that the activities these things are doing is tied back to your business and your cloud subscribers.
• Strongly encourage your cloud hardware/software platform provider to incorporate capabilities within their technology/products that will enable you to directly associate the platform or components/instances built to the business, function, purpose, subscriber and the impact a failure at that level may have on those things.
• Every monitoring event, message, page, etc. should not only communicate a message about the issue within the cloud infrastructure, but also the impact to the business and/or your subscribers. This can be as simple as stating that “200 Clients” are impacted due to this component failure. With the proper operational process and procedures, the response to this event may be initiated differently than it may have been if your event, message or page had just said “Server XYZ is down”.
• As you develop Cloud Consumer portals and interfaces, do so in such a way that they can fully understand the impact your outages, incidents, problems and maintenance activities may have on them.
• Give your cloud subscribers the capability within your portal or interface to personalize the services they get from you to their business. For example, if I subscribe to Amazon’s EC2 service, in your portal or interface allow me to map “AWS EC2 Instance 123-11-1A” to my “Inventory Management System”. Better yet, let me also add information about how this impacts my business in my language and terminology such as when this EC2 instance fails, it impacts my revenue/sales in this way for every minute/hour that it’s down. (* I’m generalizing here…I don’t know the back end of how EC2 works or if you’d ever have an “instance to subscriber function” mapping)

Cloud Consumer

• Understand how your use of cloud computing may impact your business, goals and objectives, business services and applications. Quantify this. Document this.
• While cloud computing may make the need for an IT operations staff less relevant, someone’s got to manage and monitor the cloud services you’re subscribing to actively or passively. Ensure that whomever is tasked with that role, they intimately understand the first bullet.
• Work with your Cloud Provider (service or hardware/software platform vendor) to ensure that capabilities exist so you can personalize the cloud to your business services, applications, transactions, processes, activities and how they may impact your business goals and objectives.
• Hyperic’s Cloud Status should build upon their beta (if capability exists) to enable Amazon AWS subscribers to create their own branded version (accounts?) and associate the monitoring to their business as discussed above.
• Keynote, Gomez or others should get into the market to build upon their already well established performance, availability and user experience testing and benchmarking. Keynote has recently announced Keynote Internet Testing Environment (KITE) that I’d love to see capabilities for linking my business services, applications, transactions, processes and activities to tests I run here. Follow their Web2.0 Watch blog for insight into their thinking.

What are your thoughts on BSM in the Clouds? Is it really just the same things we’ve been doing just applied to the new buzz word? What ideas or recommendations would you have for BSM concepts applied to Cloud Providers (service providers or hardware/software platform vendors) or what Cloud Consumers should expect from a subscriber perspective?

Join in the conversation!

I would like to talk to someone who is experienced with, currently uses (or has used) the Allen Systems Group (ASG) Business Service Platform (BSP) for BSM.

Please contact me via this blog post or direct.

Tks,

Doug

In an repost of an article from a couple years back, Robin Bloor provides some updated color on the state of BPM and SoA. It’s apparent that some of the other “B” buzz words have the same challenges that exist on the BSM front.

Things to ponder…

• How can these projects with such touted value to the business or IT be successfully implemented?
• Where are vendors falling short in helping “solve the organizational problems” that often cause these “B” projects to fail?
• Is throwing technology/product at the problem the best place to start?
• What should a next generation organizational structure look like? Can IT and Business organize around end-to-end business service/process delivery and support?
• How can organizations be incented, encouraged, mandated to have an end-to-end business service/process focus?
• Where success stories for BPM, SOA, BSM, BAM, etc. exist, how have these technologies been operationalized, organizations changed, workflow/process/procedure modified to reap the benefits?
• Is it foolish to think that any of these organizational challenges can ever be solved or at least minimized?
• Do we have generational issues here that will change as Baby Boomers retire and Gen X/Y/Z move up the ranks in IT and Business?

Give the post a read, I found these two very applicable to all of the things I’m seeing with BPM, BSM, BAM, etc.

Question 5: What are the most difficult steps within a BPM project – and what makes a SOA project tedious?

While I’m sure that TBSM’s use of ILOG wasn’t in the business case for IBM’s announced purchase of ILOG today, I am getting my hopes up that this is the Christmas present that anyone using TBSM is looking for.

TBSM makes extensive use of ILOG under the covers for the core of the canvas and visualization. We barely scratch the surface in capabilities (see demos below) in how its used, but it is the foundation for all of the data driven widgets used in the canvas, custom dashboards, etc.

Let’s face it, we have a long way to improve in our dashboarding and visualization capabilities. The acquisition and product convergence focus has been a significant impact on new innovation in this area within TBSM. With the upcoming release of TBSM v4.2 soon to be behind us, I hope that we can finally make investments where they are desperately needed within the data visualization, user experience, widgets, real time charting/graphing and general sex appeal areas for creating the dashboards expected of products in this price point.

TBSM 4.2 will give us a new platform to build upon, but it WILL NOT MAKE SENSE to develop other components (widgets, charts/graphs, analytics, rules, visaulziation, workflow, etc.) in the buzz word of the day (AJAX, Dojo, Python, RoR, PHP, TCR/BIRT) when we may have this ILOG portfolio at our disposal. (my opinion only, I ack the biz side of these decisions)

I strongly encourage you to voice your thoughts and requirements for improvement of TBSM’s capabilities in these areas via appropriate channels.

ILOG has a log of powerful capabilities to offer should we choose to take advantage of them. Better yet, equipping you to take advantage of them with self service designers, SDKs, APIs, etc.

For example:

ILOG Visualization Portfolio
ILOG Diagrammer (TBSM uses this)
ILOG JViews Demos
ILOG Elixer
ILOG Elixer Demos AWESOME! Look at Gauges & Dials Demo!

I hope this is the Christmas present we’ve all been waiting on for TBSM this time next year!

Compuware announced their Q1 earnings today and made a couple interesting statements on the call:

From Seeking Alpha:

As a fellow Compuware shareholder I think you should be pleased not just by our strong Q1 but by the opportunity Compuware has to help IT organizations around the world optimize their performance and get the most value out of their IT operations. The fundamental difference between us and our competition is we believe the category in which we participate of Business Service Management should be about proactive monitoring of IT performance rather than about capturing and resolving IT issues after the problems have occurred and the damage has been inflicted.

We will make the current positioning of the Business Service Management category obsolete in the years to come. So our success this quarter came through some important operational improvements implemented by our Compuware 2.0 initiative but the best is yet to come.

The weekly source for hot IT management news and gossip is the IT Management Podcast hosted by Cote’ of Redmonk and John M Willis of Zabovo. This week’s episode featured OpenNMS’s Tarus Balog.

Tarus dropped a few interesting tidbits throughout the conversation around Network Management about a couple very large IBM Tivoli Netcool clients that were moving from or complementing their existing architecture with OpenNMS. One was a large telecommunications company in Italy (Telecom Italia?) and another a very large mobile telecommunications company in Switzerland named Swisscom.

This led to some discussions around product scalability, licensing models, etc. Tarus didn’t have any specifics to share other than one requirement for OpenNMS to handle event storms of 2K-3K per second. He said they’re working through architecture approaches to ensure that their backend databases can continue to scale in ways similar to Netcool/OMNIbus’s in-memory database.

Tarus also mentioned capabilities in OpenNMS on par with what Netcool/Impact offers. I believe he called them Automations. It’d be neat to hear more on this and if they’ll have a library of data source interfaces/integrations similar to Netcool/Impact.

Everything that Tarus and the OpenNMS team does is ultimately driven back into the main code tree for all to take advantage of. The OpenNMS DevCamp kicks off in a week or two where the foundations for OpenNMS 2.0 will be worked on. This is taking place in my backyard down at GA Tech if I recall correctly.

Congrats to the OpenNMS team for your entrance into the telco space with these clients. I also really want to learn more about your Papa John’s deployment and if I heard glimpses of Business Service Management (BSM) there or if you were just using that as an example!