Salesforce Suffers: Adding to the New Year’s downtime woes, Salesforce.com also experienced major downtime this week – the site was completely down for as much as 40 minutes, affecting thousands of customers.  As Salesforce’s SaaS applications become more and more important to the enterprise (with dozens of applications/services running off the Salesforce backbone), any downtime from the giant provider can cripple customer operations and cost millions in lost revenue.  Another proving point that, sometimes, 99 percent availability just isn’t enough.

Movement in the Cloud: The cloud also saw some major shakeups this week, with EMC buying pieces of SourceLabs and Sun purchasing Q-layer, This is more proof that the cloud is here to stay, and that testing applications in and against cloud environments will only become more important in 2009.

Management Must-Dos: Network World’s Denise Dubie published an article this week looking at the “Management Must-Dos in 2009,” with ITIL, specifically process, being one of the top concerns.  We couldn’t agree more.

The downtime winner of the week is Twitter! Sporting a new downtime image, Twitter experienced a significant amount of downtime this week. SMS delivery problems are speculated to be a reason for the downtime, since the Twitter Status blog did not indicate the reason for the database efforts.

ITIL v3 recommends using Key Performance Indicators (KPIs) to measure the effectiveness and efficiently of the ITIL process. The problem is ITIL is not a regulatory for measuring services. The business values of the KPI service are utility and warranty. CIO provides a seven step improvement process to identify KPIs for a service. The steps include defining what you what you should and can measure; gathering, processing, then analyzing the data; presenting the information and finally implementing corrective action.

According to Michael Keen, Director & Senior Solutions Architect for the Enterprise Architecture Group at Alliance Technologies, enterprise architecture is just as important as IT governance. He says that enterprise architecture shows how all the components of the enterprise are related and it provides a framework for supporting and automating business processes, especially with change. Keen claims that clients should not ‘set and forget’ enterprise architecture because it is a dynamic, disciplined and ongoing process. If enterprise architecture is properly envisioned and implemented, it will enable IT orgs to respond to change rapidly.

Mike Kavis, CTO of Kavis Technology Consulting, shares his presentation on ‘Managing Change in Business Transformation Environments.’ He discusses why transformational IT initiatives like BPM & SOA fail and offers on how to prevent failures. Mike also reveals his thoughts on the future of IT and how a well ran and planned enterprise architecture is a key aspect of a successful implementation of transformational change.

Demian Entrekin, founder of on-demand PPM provider Innotas, shared his 10 predictions for the Future of SaaS and On-Demand Software Applications. He says the key to driving adoption will be better product integration and alliance with other SaaS companies. Among the 10 trends include software without borders, more product alliances and shifting more energy to market strategy.

Peter Farquharson, technology integration service manager for the City of Saskatoon discusses ITIL implementation and change management for his organization. He points out that although the city is not doing a lot to promote ITIL, he is taking measures with change management systems. Farquharson wants to see a “citywide management system with a Web-based front end that allows citizens to ask for a pot hole fix and have it happen as seamlessly as the city can now respond to technology change requests or incidents.”

A new report from Gartner predicts a ‘worldwide enterprise SaaS market will surpass $6.4 billion this year, up 27% from 2007.” Sharon Mertz, the research director at Gartner, recognized the popularity and growth of the on-demand deployment model. The factors for deployment are the current economic state, better broadband and a need to ‘rapidly deploy software to meet a specific business need.’

In a recent survey conducted by IDC and sponsored by HP, a majority of IT organizations at large companies worldwide see reduced IT costs as a potential outcome of implementing ITIL. The survey says more than one-quarter of organizations in the Americas are already implementing ITIL and ‘over 60% indicated they plan to bring the best practices in house within the next one to six months.’

John Michelsen, iTKO’s co-founder and “Chief Scientist” shared his three step process for successful SOA testing: development builds the testing framework, development builds initial test cases, and QA enhances the test cases. Based on his customer base, Michelsen said one of the characteristics of a successful SOA implementation is integrated testing in the development cycle.

More detail from Kurt’s responses and answers to similar questions are spotlighted in StackSafe’s Best Practice Guide: Application and Infrastructure Upgrade with focus on pre-production release and testing. You can download the entire guide for free.

StackSafe: How did you get the data for the recent report of application upgrades?  How many companies did you research?

Kurt Milne: We used data and analysis from multiple IT Process Institute studies to develop our summary of best practices related to the application and infrastructure upgrades. Our study methodology is focused on identifying practices that have a measurable impact on a range of key IT operating performance measures.  We conducted three different web-based surveys and collected data from over 1000 companies that was used to identify a range of practices that best predict top performance across the software development and production release lifecycle.  Our study of Change, Configuration and Release practices was referenced in the footnotes of the best practice guide.  Free summaries of all our studies are available on the whitepaper section of www.itpi.org.

StackSafe: Can you explain the AIM framework and its impact on organizations?

Kurt Milne: We put our recommended practices into a staged framework that made sense to the serial CIO on our study team.  Most organizations want a simple way to scale into new practices to help build momentum for organizational change.  Using a staged approach for implementing application and upgrade best practices across the development and release lifecycle helps identify where to start, and also build a vision for a top-performing future state. The Adjust level practices identify initial changes that should yield immediate measurable results.  The Improve level practices build on early success and identify additional ways to increase performance.  The Master level practices round out our recommendations and should help organizations achieve operational excellence and top levels of performance. 

StackSafe: How do process enhancers like ITIL and COBIT fit into your recommendations for best practices?

Kurt Milne: We develop our data collection surveys based on what we hear during IT executive interviews.  Many of the practices identified during those interviews come from the study participant’s use of frameworks such as ITIL and COBIT.  We don’t set out to test those frameworks specifically.  But our approach often identifies the handful of practices, out of literally hundreds of practices outlined in these frameworks, that are shown to have a broad impact on the sample of companies in our studies.  That doesn’t mean that the other practices in these frameworks should not be considered or used by IT groups.  But our approach identifies those practices that are commonly used by organizations with higher levels of performance.

StackSafe: How can an IT Operations teams measure the success of their efforts in configuration management and application upgrades?

Kurt Milne: We recommend two types of measures. 

• Process measures that verify practices and controls are implemented. 

For each area of the best practice guide, we offer several measures that help indicate whether those practices are being followed.  For example, in the configuration management section, we suggest measuring the number of systems that are verified to match golden build, and the time to rebuild a system from bare metal.

• Outcomes measures that indicate the operational performance of the organization. 

For each area of the best practice guide, we show measures that are likely to improve as a result of implementing the practices, and determined from our empirical studies.  For example, in the section related to linking change request to business and infrastructure context, our studies have shown that the use of recommend practices strongly correlate with lower release rollback rate, and reduced configuration drift.  The combination of process and outcomes measures helps IT organizations simultaneously focus on the practices and their impact.

StackSafe: How does a company’s culture impact their ability to effectively change and manage their current processes?

Kurt Milne: Having a process culture is one of the strongest predictors of top performance we have uncovered in our studies.  It makes sense logically, that if you drop best practices into an IT shop that doesn’t have a history or culture of following documented procedures, then those practices will not be consistently followed. If best practices are not consistently followed, then the desired results are uncertain.  My opinion is that the process culture is more important than the specific practices.  In other words, consistently following a sub-optimal process is likely to give better results than inconsistently following a best practice.  This is especially true in the control areas of access management, change management, and configuration management.  Variability in these processes introduces unknown levels of operation and security risk.

StackSafe: What steps can companies take to ensure that infrastructure and application upgrades enhance business capabilities and minimize disruption?

Kurt Milne: The longer I study IT operations, the more I am convinced there is a basic set of operational practices that should be implemented in some form, by all IT organizations.  This best practice guide offers a simple set of practices, in a staged format, that IT shops should consider for optimal performance.

Martin over at Blade Watch (congratulations on two years!) asks whether or not IT and business are set to become more aligned. He says, “It’s not enough to do IT, we need the team to understand the infrastructure and the application to know what happens if server17 fails, what if feed4 breaks, how to fix or co-ordinate issues, how to manage and continually improve IT in terms of service and delivery.”

We had the opportunity to speak with the IT Skeptic and John Willis in the past. It seems we also helped to set up a fine connection between the two as both posted brief Q&A’s with one another. Check it out for more information on ITIL, ITSMF, and what makes operations management so exciting.

Does Google really have 99.9% uptime? According to TechCrunch, Google announced that the enterprise version of Gmail comes with a 99.9 percent uptime guarantee. Google is extending that guarantee for enterprise customers to Google Calendar, Google Docs, Google Sites, and Google Talk. Radicati Group conducted research to compare services like Microsoft Exchange and found that Google and Gmail were four times as reliable.

Lydia Leong blogs for Gartner and provides a compelling case for cloud computing for IT Operations teams. Lydia says that IT Operations will soon be automated, but that many enterprises are still lacking the scalability they need to do that effectively. Even so, “the shift to global-class computing, and massively scalable infrastructure, represents a fundamental shift in the skill sets that will be valued in IT Operations.”

E-commerce websites are going to be under a watchful eye as the holiday season approaches. Uptime becomes even more critical this year as retailers are wondering if consumers are going to shop online. If so, then downtime certainly cannot happen, which means companies need to effectively plan for changes, upgrades, and scalability issues BEFORE the shopping rush hits.

Speaking of downtime, Amazon just made some announcements for EC2, including an SLA that states, “…each region (right now there’s one, but more are coming) will have 99.95% availability.” That’s a hefty promise, and hopefully this can be a success story for Amazon, rather than another cause for criticism.

Microsoft is talking a lot about virtualization. A recent post on the Virtualization Product team blog from a marketing manager for Windows Virtualization at HP cites a few reasons why virtualization is a key component to their strategy. Business continuity and availability (no more downtime!), cost savings, and going green are all reasons to consider virtualization. Microsoft and HP are also working together to develop a single tool to manage physical and virtual environments. Now, if we could only get Microsoft and other companies to effectively test their changes before making offering upgrades…

Speaking of Microsoft upgrades… Microsoft is begging people to upgrade to Vista and IE8, etc. They want you to consider the risks of not upgrading, while customers are worried about the savings involved with not moving to Vista and the latest version of Office. What’s the risk of not moving? Gartner predicts that vendors won’t support new applications running on XP starting in 2010.

Is Robert Scoble saying enterprise software is sexy now? According to Michael Krigsman, he is. We discussed our thoughts on Robert’s highly controversial post last December, and through a few new posts, it appears that “Robert’s position on enterprise software has deepened since last year. At that time, he seemed to think software value could be defined by cool features and nice looks alone; Robert now recognizes that substantive ROI is the true measure of technology.”

The IT Skeptic has written a book about ITSM, saying that “Vendors, ITIL, COBIT, and CMMI are not safe in Real ITSM”. Check out John Willis’ initial thoughts on it and make sure to check out the book. It should certainly be entertaining, to say the least.

In a recent links list, we noted that the Configuration Management Database (CMDB) is becoming a more common component of companies’ IT infrastructure. In just two years, according to Enterprise Management Associates (EMA), the number of companies already implementing a that have reported completion of a CMDB has have seen completion rates jump from 9 percent to 33 percent. Given the complex nature of the CMDB effort, this is a significant jump.

This movement tends to confirm the findings that we recently noted in a report on Change Management Maturity. In short, companies notice significant business benefits – like a smoother change management process and fewer production problems – when they automate change management and adopt a standard such as ITIL. ITIL guidelines stress the importance of establishing a CMDB to identify and control the configuration items (all the technical ’stuff’) that exists in the infrastructure. Automating change management requires a clear understanding of what you have in the infrastructure, where it is located, and what state it is in.

The article from bITa planet quotes a similar finding from EMA:

EMA’s research has shown that 60 percent of outages have change as their root cause, and the CMDB in service of the ITIL Service Support discipline that includes change management provides the holistic view of the IT infrastructure necessary to make good change decisions.

So the benefits would seem to be there, but the problem is often justifying the cost of a CMDB effort. Especially in rough financial times, IT projects will be judged more favorably if they demonstrate ROI than if they are simply advancing IT capabilities or introducing less tangible benefits like flexibility or risk avoidance.

In the bITa planet article, EMA advises “wait as long as possible to start hard ROI for your CMDB. A lot of people start too early and are disenchanted with what they get out.”

This may be easier said than done. CMDB’s are not deployed to production overnight. Frankly, because CMDBs must reflect the changing state of the infrastructure, a CMDB may never actually be ‘done’. There are other methods to avoid downtime, like better testing and more complete testing environments. A CMDB can provide an organization with a myriad of benefits, but the project to build a CMDB is going to need to be more closely mapped to ROI than ever. Companies must look at CMDBs as strategic investments - developing a CMDB now won’t deliver short-term ROI, but delaying CMDB development also pushes back the time to ROI.

On that note, Michael Krigsman is urging companies to “stop brain-dead software testing” and take the time to adequately test changes and upgrades. He cites errors, inadequacies, and oversight that lead to errors that often cause unplanned downtime. For example, clothing retailer J.Crew recently had serious customer relationship problems when it deployed poorly tested software. Effective change management and testing are keys to IT project success.

Tarry Singh has some data center predictions for 2009. More details to come, but an example of a few issues that were relevant to IT’s About Uptime: consumerization, industrialization (data center, practices, etc.), operational excellence within IT: SCM, sales automation, marketing automation, bid management, project management etc.

In a survey of 316 senior financial industry executives, SAS found that most organizations don’t have a high-priority culture around enterprise risk management. The results show that executives are reluctant to assume the responsibilities of risk management. One priority IT staff needs to improve is reducing IT project failures. Three steps to build an IT success culture are establish priorities, invest in activities and demand conformance and reward improvement.

Is there such thing as ITIL compliance? IT Skeptic tackles this topic and states that there is actually no such thing as an ITIL “non-compliant” software tool. In fact, any software tool can support ITIL. Yet, many people question how particular software supports ITIL processes, so he provides questions to test tools for ITIL “compliance”.