Home > Blogs > VMware Operations Transformation Services > Monthly Archives: April 2013

Monthly Archives: April 2013

VMware #CloudOps Friday Reading List: Orchestration and Automation in the Cloud Era

Shift your thinking from task automation to service automation.

There are two primary automation strategies. Shift your thinking to service automation in order realize the full potential of SDDC and hybrid cloud solutions.

Automating IT Services Is Not As Hard As You Think by Edward Newman
Your 1st automated service provisioning effort should be a complete service that has big impact. Not automated provisioning of multiple VMs with manual storage and network deployment.

The Importance of Automation in Cloud Computing by James Bond
Good step by step summary of what happens during automated provisioning. Great advice to ask your cloud service provider for a detailed flow diagram or process description of what their automated workflow system is doing.

Making Change Management Your Cloud Computing and DevOps On-Ramp by Hank Marquis
Use your standard change process as automation on-ramp. This article includes a great list of specific measures that gauge maturity of your standard changes.

5 Benefits of Cloud Management by Tony Lucas
Automation can help IT organizations turn traditional hardware into a cloud platform, streamlining deployment while managing and scaling resources appropriately. Use management platform metering capabilities to better understand service costing.

Follow us on Twitter at @VMwareCloudOps for future updates, and join the conversation using the #CloudOps and #SDDC hashtags.

IT Automation Roles Depend on Service Delivery Strategy

Many of the agility and cost reduction benefits realized by deploying standardized and virtualized infrastructure (compute, network and storage) come from automating management processes.

But those benefits don’t come (forgive the pun) automatically. Automation is an approach to achieve a goal. And your reasons for deploying automation greatly influence both the tactics you follow, and their  impact the people, their roles, and skills required to function in a more automated environment.

Strategy Impact on the Automation Manager

I recently spoke with the IT Director of a large multinational bank. As a leader in the IT Operations organization, he is creating a new ‘Automation Manager’ function.

Until last year, he’d been working to get key bank ITIL processes to level 3 or in some cases level 4 maturity, and he’d achieved a state of advanced operations which benchmarked well against peers. As part of that effort, he’d been deploying a group of ITSM process owners in a matrix organization structure with central process owners working part-time placed in various business units. That approach worked well.

Now, though, his focus has shifted from process maturity to process automation. Yes, the department had a foundation of mature and consistent process, created by expert staff. But to get to next level of efficiency, he told me, they needed to increase automation. To help with that, the director created a new role: the Automation Manager.

This wasn’t the first time that I’d heard about such a job. This particular conversation, however, highlighted various questions shaping the new role:

  1. What type of automation? There are different types of automation: powershell script, cron job, workflow tool, policy based orchestration, configuration automation. There are the different activities that can be automated: provisioning, maintenance, scaling resources, proactive incident response. And there are different degrees of automation: automating just a few actions, partical workflows, or going end-to-end.
  2. What is the job scope? Automation has a lifecycle: intake, classification, resources, version control, tracking benefits. What part of it does an Automation Manger own?
  3. Where does the role fit in the organization?  An Automation champion ideally owns overall program, but is it also someone to whom others can turn for technical automation advice? The IT Director’s idea is to create a central role. But where does it fit? The ITSM process group? The technology team?

All of these questions had me stepping back and thinking at a higher level. The answers depend more on the strategic goals, and less on the tactics.. Given that, I think two primary automation strategies  frame decisions about the Automation Manager role.

  1. Task Automation. Here, automation helps staff do existing work more, faster, better. With this strategy, the same people continue doing pretty much the same admin jobs they did before. The new automation manager becomes an overlay function that helps each admin use new tools to turbo boost their existing work.
  2. Service Automation. Moving from a technology or infrastructure focus to a service orientation requires more standardization and automation. This strategy is about automating new processes that didn’t exist before. In many cases, automation enables workflow that wasn’t even possible in a manual process approach. The automation capabilities support admin roles that may be largely new, or may be a combination of previously separate roles.

The Value Proposition Tradeoff

One way to think about these options is by revisiting the basic tradeoff between speed and efficiency, versus customization.

Consider the analogy of the clothing business. If your customers want affordable clothes now, you can offer “off the rack”  in standard sizes. If they don’t mind paying more and waiting, you can offer  something custom-made. They’re simply different value propositions.

In the clothing business, there are uses for automation that help tailors deliver custom work faster, with fewer errors. But automation can also be deployed as part of a strategy to create goods in standard sizes on a massive scale. They are two different automation strategies.  The role of the automation manager is either help tailors improve custom work, or build a factory to mass produce standard sizes.

Similarly IT can deploy automation to help execute existing work faster, better, and with less effort. Or IT can use automation to deliver highly standardized services at scale. Either way, if you’re clear about the strategy, the details about the Automation Manager role will come into focus.

For one, helping existing staff do more, better, faster, requires a role largely focused on implementing tools, training users, and offering support. For the other, building an IT factory that delivers standard “off the rack” services at scale, requires a role that is a process and systems engineer who builds and maintains factory robots that do the work.

Clearly, the IT Director I spoke of has a more, faster, better automation strategy.  What strategy do you have?

What can IT admins do to better position themselves for their new responsibilities in the cloud era? Find out by joining a live Twitter #CloudOpsChat, on “The Changing Role of the IT Admin” – Thursday, April 25th at 11am PT.

We’ll address questions such as:

  1. How does increasing automation change the IT admin job? #CloudOpsChat
  2. Is increasing automation and virtualization good or bad for your career? #CloudOpsChat
  3. Do abstraction and better tools decrease the need for deep expertise? #CloudOpsChat
  4. Does a cloud admin need programming skills? #CloudOpsChat
  5. What skills are needed for scripting vs. automation and orchestration? #CloudOpsChat
  6. What specific automation skills do IT admins need today in order to meet the demand for virtualization and cloud technologies? #CloudOpsChat

Follow us on Twitter at @VMwareCloudOps for future updates, and join the conversation using the #CloudOps and #SDDC hashtags.

Recap: 5 Key Steps to Effective IT Operations in a Hybrid World

In a recent webinar, VMware’s Kurt Milne, and Kevin Lees, Principle Architect for Cloud Operations, discussed how to optimize IT operations in a hybrid world.

Users and business managers can now easily work around IT to procure a wide range of services from vendors on-demand, and get billed based on usage. So what can, or should, IT organizations do?

Here’s a quick summary of the key points from the discussion. Kurt and Kevin provide recommendations and solutions to avoid shadow IT activity, and to make cloud era IT operations as effective as possible.

Problem #1: Your business wants fast AND easy.

As IT evolves, users increasingly want immediate access to central IT resources and services. They’re adding agility requirements to the cost, performance, and service quality requirements that have traditionally driven IT strategy. How do you respond?

Solution: IT needs to shift away from building custom resource configurations for each application and take a more standardized approach – building a factory that can deliver self-service, on-demand access to a portfolio of “off-the-rack” services. Users will get the control and speed they want if they’ll agree to standard service configurations. They can get fast and reliable, but not custom.

Problem #2: Transparent Service Deployment

Increasing deployment options (physical, virtual, private and public cloud) are pressuring IT to offer solutions that are, as Kurt puts it, “fit for purpose,” in terms of duration, cost, quality, performance and business context. This suggests a need to  shift from a custom systems view, to a standardized portfolio view of both workloads and compute environments.

Solution: Deploy systems to an environment that is a best fit for a range of criteria. Collect attribute data during service requests, then use a policy-based system to make workload placement decisions. Enable automated provisioning that can cross boundaries and work with physical, virtual, and multiple cloud environments. Your business really cares about the cost, risk and quality of service. But if IT is responsible and accountable for performance, service quality, support, as well as security and compliance, it also needs to stay in control of workload placement and provisioning decisions.

Problem #3: “Run Time” Complexity

Modern application topologies and compute stacks are more complicated, and often more dynamic, than ever. As a result, many IT organizations are re-evaluating resiliency strategies from the ground up. In the past, we improved service quality by carefully controlling changes to a static environment. What do we do today?

Solution: In the cloud era, it’s better to assume that change will happen, and build systems to monitor and automatically respond to likely conditions, in order to deliver high quality of service. We need to shift from reactive incident response to proactive detection and preservation of services before outages occur. This sounds like a utopian vision. But it’s backed up by real solutions that let you start making powerful changes now – by leveraging a single platform for work across environments and across pre-production and production lifecycle phases.

Tip 1) Focus on performance, capacity, monitoring and analytics. Leverage intelligent analytics to identify conditions that predict pending disruption. Use the same capabilities to accelerate root cause analysis, determine the source of the pending disruption, then fix it before services are impacted.

Tip 2) Don’t forget to plan for new roles and skillsets needed in a policy driven and automated environment.  (Attend April 25th Twitter chat on Changing IT Admin Roles hosted by Puppet Labs CTO Nigel Kersten and VMUG IT founder Andrea Mauro)

Problem #4: Workflow Variability

Ad hoc operations is a weak foundation for automation. Highly variable procedures for building or maintaining complex systems are neither repeatable nor deterministic. Exceptions drive up the manual response costs. And exceptions also reduce service quality.

Solution: Standardization and policy based automation is key for enabling a shift  away from variable and unpredictable “Artisanal IT” to predictable and reliable “Industrial IT”. Adopt a policy-based control strategy to guide workflow and automation. Start by setting basic policies, then look for process and run book procedures that require exceptions. Then eliminate the cause of those exceptions and standardize activities to fit policy.

Problem #5: Alternative providers have clearly marked prices

IT does not clearly price services, for both historical and structural reasons. It is certainly easier to fund IT based on project or flat rate business unit allocation. But 3rd party cloud service providers offer service- and usage-based pricing, which is changing the expectations of users and IT funders. IT won’t be the preferred service provider for a business if its services are not priced for comparison. It’s time to act.

Solution: IT needs to enhance IT financial management. It needs to put systems in place that provide a service-based view of costs. It needs to fund operations at least in part based on services delivered. And it needs to offer pricing information at the point where services are ordered to enable cost comparison.

The good news is that IT has an insider view of business objectives and the unique needs of users. With improved transparency, and a portfolio of services that meet the needs of your business, IT has unfair advantage over 3rd party service providers.

Round Up: By acknowledging and addressing these problems your business can offer effective IT operations in a hybrid world, and will possess the advantages needed to be the preferred service provider for your users and business managers..

For solutions that enable the approach that Kurt and Kevin were talking about, check out:

vCloud Director Learn more
vCloud Automation Center Learn more
vCenter Operations Management Suite Learn more
vFabric Application Director Learn more
vCenter Chargeback Manager Learn more
IT Business Management Suite Learn more

Want to learn more? Check out the webinar replay and follow us on Twitter at @VMwareCloudOps for future updates! Join the conversation using the #CloudOps and #SDDC hashtags.


VMware #CloudOps Friday Reading Topic – Service Definition Process

Increasingly, we see the service definition process as a key dependency in the success of a hybrid cloud or SDDC strategy. Standardization of service offerings (and thus configurations, as well as management and maintenance processes) is key to simultaneously achieving agility and efficiency benefits.

Here are some interesting Friday reads related to standardization and the service definition process.

Putting The Service Back In “as-a-Service” by CloudTweaks
Pete Chadwick offers advice on how to uilize a service-oriented approach to ensure the business can easily access and rapidly deploy what it needs.

Preventing Epidemics in Cloud Architectures by Gordon Haff
Gordon digs into a recent presentation by Netflix’s ubiquitous Adrian Cockroft. Understand the tension between the benefits of standardized services, and the inherent weakness of a homogenous environment.

ITSM Goodness: How To Up Your IT Service Management Game In 7 Steps by Barclay Rae
To achieve ITSMGoodness – start by listening to customers, and structure services based on business outcomes. Services trump SLAs.  Good perspective from visionary Barclay Rae!

Service Initiation: Understanding the People and Process Behind the Portal by David Crane and Kurt Milne
In VMware’s CloudOps operating model, Service Definition is one part of the multi-part service initiation process.  Listen to this webcast to understand how these four processes fit together.

Follow us on Twitter at @VMwareCloudOps for future updates, and join the conversation using the #CloudOps and #SDDC hashtags.

The Changing Role of the IT Admin – Join Us For #CloudOpsChat!

Cloud-based resources, automation, virtual everything, DevOps tool chain – are all causing change to the traditional role of admin.

In some of our recent blog posts we’ve discussed whether new technologies are good for careers in infrastructure and Ops, as well as the variety of ways IT can transform itself to optimize results.

What can IT admins do to better position themselves for the new responsibilities in the cloud era?   Find out by joining a live  twitter chat #CloudOpsChat on “The Changing Role of the IT Admin” – Thursday, April 25th at 11am PT.

The event will be co-hosted by two very knowledgeable experts on this topic:

We’ll answer questions:

  • Is increasing automation and virtualization good or bad for your career?
  • Does abstraction and better tools decrease the need for deep expertise?
  • Does a cloud admin need programming skills?
  • Do you think a cloud admin is a generalist with a little expertise about compute, storage, and networking?
  • Or is a cloud admin an expert in cloud architectures?
  • What are your thoughts on scripting vs. automation and orchestration?

…and more.

Here’s how to participate in #CloudOpsChat:

  • Follow the #CloudOpsChat hashtag (via TweetChat, TweetGrid, TweetDeck, or another Twitter client) and watch the real-time stream.
  • On Thursday, April 25th at 11am, @VMwareCloudOps will pose a few questions using the #CloudOpsChat hashtag to get the conversation rolling.
  • Tag your tweets with the #CloudOpsChat hashtag. @reply other participants and react to their questions, comments, thoughts via #CloudOpsChat. Engage with each other!
  • #CloudOpsChat should last about an hour.

In the meantime, feel free to tweet to or at us (@VMwareCloudOps) with any questions you may have. We look forward to seeing you in the stream!

Understanding the People and Process Behind the Portal – Register for Webinar This Wednesday!

One great way for IT to adequately address the needs of increasingly tech-savvy and impatient users is by implementing self-service on demand service provisioning.

David Crane, a Cloud Operations Consulting Architect and Kurt Milne, the Director of Cloud Ops Programs at VMware are hosting a webinar, tomorrow, April 10th, at 8am PT on “Service Initiation: Understanding the People and Process Behind the Portal.

During the webcast, they’ll discuss key concepts related to being a service broker in a hybrid world, as well as a cloud-optimized approach for managing a service pipeline. David and Kurt will specifically highlight four interesting IT roles that will help to optimize this approach.

Join us as we dig into this important issue and more – register for tomorrow’s webinar now!

In the meantime, feel free to tweet at @VMwareCloudOps on Twitter and join the conversation using the #CloudOps and #SDDC hashtags!

Would All You Users Please Just Behave!

By: Ed Hoppitt

A significant slice of the problems that those of us running IT are dealing with today comes down to badly behaved users. In fact I spoke about the issue of badly behaved users in my presentation at VMworld last year.

What I mean by “badly behaved” is that users are becoming more and more impatient and demanding.  They are used to having their own way in the consumer world of technology. And now they’re bringing their impulsive, demanding and impatient behavior to the workplace.

I get it.  Time is money.  Users and business managers who fund IT are impatient because their customers are impatient.

But as IT professionals, we are now dealing with agility and responsiveness requirements as much or more than efficiently provisioning and managing the storage/compute/whatever infrastructure resource we think we are supposed to be focused on.

I’m not just talking about the issue of provisioning passwords for someone – that’s the easy bit. The real issue is created when you add the magic of ‘Self Service’ to the mix. How long does a user have to wait for an application on their iPad? About as long as it takes to download from the App Store. They expect a similar experience from corporate IT portal.

Do they forget the times when getting a new application at work meant installing from multiple tapes/disks/CD-ROM’s? You had to  fill  in a request form. Then someone posted you the media on which your application existed. Then you installed it (not of course forgetting the 3 or 4 failed installations that you’d try before you got it to work). I’m going off topic.

My point is that as part of deploying and managing IT, we need to realize that the expectations of users have changed significantly.  Unless you factor rising user expectations into your IT strategy, you will fail at the most important part – ensuring that people are having a good experience.

What can an IT Ops professional do to manage rising expectations?

  • For starters – when someone presses a button in a self-service interface, give him or her confirmation that what they expect to happen is happening. Don’t leave people wondering as to what’s going on.  The same is true when it comes to provisioning – the longer it takes, the more inclined people will be to over-order, over-anticipate and under utilize.
  • Also, give people the ability to take resources as easily as they can give them back. If someone stops using something, encourage them to return those resources, credit it back to their ‘credit card’ of IT services they can purchase…. Turn self-service into an incentive for good behavior.
  • Focus effort on communicating value back to your business and to your users – or people don’t see and realize what you are doing.   Tell people what these changes will mean and sell them the benefits. So often IT does what it does inside the IT microcosm without learning from Marketing colleagues about a little self-promotion.
  • And if you really want to impress users, ask them what they want. As technologist and engineers we sometimes forget this important step.

As an IT admin, you can also directly ask users what features they want to see from IT or what would be useful to them. Reinforcing the idea that IT is there to make their lives easier can help persuade users to follow correct IT practices.

It’s not hard to realize that users work for the people that pay IT’s bills, therefore….

Happy users = Happy IT

This is a modified post from Ed Hoppitt’s personal blog, Gathering Clouds.