I’ll start with this truth: the truth about software vendors and the push to enhance. I’ve been saying for some time, “Vendors think buyers want more features, when really they don't.” Things are already complex (see more by Greg Ness, on twitter as Archimedius). A recent story, another writer: here.
"Executives overwhelmingly believe that the strongest differentiator is when vendors understand the buyer's business and prescribe tailored solutions," Santucci writes in his article.
This shows there’s something myopic about the IT industry. Vendors – backed by leading analysts such as Forrester, Gartner - keep striving to create a technical edge, generally leaving it to the customer to work out how to deploy it in an integrated way with existing tools and processes. A current theme of mine is how analysts keep writing reports about how great the new infrastructure technology is and we should all be diving into dynamic IT and cloud computing.
Impact of next-gen IT
A recent familiar-themed report, from Forrester, titled “Increase Resiliency And Flexibility With Virtualization” published March 2010, has many positive comments on the value of increasing virtualisation, but hardly mentions the importance of IT automation to improve processes.
The article does note the emerging need for a couple of key process improvements:
* policy-based automation – but doesn’t say which policies or how they can be automated
And then
* develop operational efficiencies – again without any advice on which or how
This is just one of a slew of articles telling us to virtualise for resiliency & scalability with barely a nod to the CRITICAL importance of operational processes. We have clients starting to stress about how their IT ops processes – barely able now to cope with BAU and keep a process maturity cycle active – will handle the looming business demand for a switch from legacy IT to ITaaS, private cloud, dynamic infrastructures, etc. Everyone, other than IT Ops, seems to assume the legacy is easy to replace.
Put it this way – IT operations are the nervous system of IT delivery. If we ramp up the demand, the "consciousness" of business services, without upgrading the nervous system there will be a breakdown, just as in organisms that are stressed beyond their usual operating capacity. I’ll borrow more of this analogy from another writer – put the mind of an evolved human being into the brain of say a chimpanzee, what will you get? A smart chimp? No, an instant nervous breakdown.
Some other key findings in this and echoed in other reports - the top virtualisation drivers:
1. Create a more efficient shared IT infrastructure
2. Disaster recovery & business continuity
3. Improve IT infrastructure manageability & flexibility
However another very common finding is the virtualisation threshold that prevents these drivers from being fully executed. We deal with this issue all the time. Why is it hard to virtualise past the typical 30-40% of servers barrier? That’s simple.
The IT Operations challenge – crossing the threshold
At this level the challenge is to bring over the mission-critical apps and services, around which there is many years of built up legacy infrastructure. Virtualising them means reworking all the legacy processes, which means all of IT operations has to learn new skills, new management tools, and collaborate in ways that are new and appear risky.
Our clients tell us, as noted above, that the barrier to re-working infrastructure processes is BAU. The often-repeated cry is “my people are flat out keeping up with the current change activity and running operations; we have no time to re-work critical processes for new infrastructure.” The prospect of replacing legacy infrastructure, with all it’s built-up complexity, with something simpler , dynamic and more resilient is clearly appealing, but also daunting for IT operations.
For most IT organizations there is a lot of inter-dependency in current IT Service Operations for critical IT services. Between systems, and staff & vendors. All the low apples that can be virtualized easily have been done, which gets us to the 35-40% virtualization threshold commonly reported.
However, there’s much more risk in not changing. As businesses demand the benefits of a virtual dynamic infrastructure, for faster scalability and greater resiliency at less cost, IT operations will have to seriously raise the game. What has worked for fixed legacy network systems for 20 years must eventually be replaced.
Why do we care?
How can we help, what’s our angle? Well we happen to know a lot about improving processes – it’s been our core business for over 10 years. We know a lot about the need for increased integration. This shows up typically as a need for integrated configuration management tools and data, or as a need to radically improve release deployment processes. Or sometimes as a need to get on top of the “event storms” that generate from early attempts at automating problem diagnosis & analysis.
These are early symptoms of an IT nervous system that’s pushed to a limit, things start to break. Even operational change management for many relies on immature and outage-risky processes. Not because they’re difficult – usually they’re very simple – it’s just the volume & variety of changes, in a risk-averse operational culture with a complex infrastructure, has prevented effective automation. What we’ve developed is an architecture approach as a roadmap to building the automation needed for a dynamic and highly virtualised infrastructure. This ensures that the constraints of existing infrastructure, with the benefits, are absorbed, while tools and data are re-used wherever possible.
The goal is simple – to enable more resilient and scalable processes requires a re-work of IT Service Operations, represented by various problem symptoms that recur. Process maturity enables the scalability and efficiency gains from increasing virtualization, which now needs to fully embrace mission-critical and legacy-dependent applications.
If the benefits are so good, what are the barriers?
This is easy to summarise, from talking to many clients involved in meeting business demands. Several key points are already noted:
• Licensing, still a largely inflexible last-century model from most vendors
• Legacy infrastructure runs mission-critical systems, which will be sticky for many years yet
• Enterprise software upgrades and releases are inter-twined with myriad internal projects
• IT operations have legacy-based processes evolved over decades, they’re also sticky
• The CIO might be aware of the need for change but operations management, wedded to a “keep-the-lights-on” culture, tends to resist, which is completely rational
So there’s a mix of technology, process, education and culture issues that need to be addressed and planned, allowing the changes to be understood and built into an outcomes-focused maturity cycle. From our perspective this is Process 101 and the only thing new is some of the technology.
What next?
To make this higher ground?
The eco-system is already in place. Where marketing hype is processed as technology promise, this creates the ideal opportunity for smart business-process specialists, like us, to ensure the promise actually works, that the infrastructure investment provides an ROI, and that the IT operations processes can scale to support the growing business demands from IT services. Everything about the technology is focused on the right business issues – flexibility, scalability and resilience for a new utility-based computing services paradigm. It’s just that getting there will need a major and complete re-think of IT operational processes. Really that should be BAU.
Vendors - keep on vending, improvers - keep on improving. World, keep on changing, ‘cause it won’t be too long (TY Stevie) … for the truth to be shown, in all it’s shrunken glory.
David Gandar
MD Delta Software
ITvirtual
<< News < Delta Software sponsors a new fund-raising challenge | Real-world processes for a virtual age >
post a comment
Post Comment