As we illustrated last month, Toyota used the principles of Lean Thinking to propel itself from obscure manufacturer to global auto-making powerhouse in the 1950s and 60s. Today, data-centric firms are using those same timeless tenets to their corporate IT functions using software-defined infrastructure (SDI) to streamline service delivery and eliminate muda (the Japanese word for waste). Under optimal conditions, SDI capitalizes on the Lean principle of flow to deliver computing assets in a single fluid motion from start to finish. But eventually, all firms encounter situations where flow just isn’t possible. That’s when SDI shows its versatility by applying the Lean principle of pull to provision IT infrastructure quickly and efficiently.
According to James Womack, author of the landmark book, Lean Thinking, flow can be described as the transformation of raw material into finished good in one continuous forward-moving operation from beginning to end with no stoppages, rework, or backflows. If that doesn’t sound anything like the way IT operates in your company, don’t feel bad – you’re in the majority. Most corporate technology departments deliver IT services via batch-and-queue — the archenemy of Lean.
Provisioning resources the hard way
For example, a line of business (LOB) manager needs a new cloud instance for an upcoming seasonal marketing campaign, complete with Microsoft office and the latest version the firm’s custom marketing application. The IT manager immediately divides and assigns the required tasks among his department’s platoon of IT specialists. The procurement specialist specs out and acquires the necessary hardware. An administrator decides on the most appropriate server operating system and installs it. She throws it over the wall to an analyst who installs Microsoft office, who then passes the instance to the delivery team to install the custom application the marketing analysts will use.
After all that installation and configuration, a quality analyst tests the installation to ensure everything functions properly. If QA finds any problems with the finished product, back it goes for remedial work and possibly another trip through the entire chain. Depending on the company and IT’s backlog at the time, this could take anywhere from six days to six months. Until then, the poor marketing team experiences one of the most frustrating forms of operational muda: wait time.
SDI for lean process improvement
The process of provisioning new compute resources illustrated above is typical of the average global corporation. It’s also about as lean as a Big Mac with a large fries. And while the muda-ridden value chain would appall Lean expert, James Womack, it wouldn’t surprise him. Most technology departments naturally gravitate toward the batch-and-queue mode because IT is a department of technological specialists.
Experienced technology professionals are in high demand and short supply. Therefore, CIOs must hire professionals with specialized skill sets that seldom overlap. Once their team is in place, work is doled out to each according to his specialty so that no single worker is involved with the deliverable from start to finish. The very structure of the modern IT department has created a disconnect, rendering the concept of Lean flow nearly impossible to achieve.
The lean principle of pull
Lean has rule of thumb for situations like this: “Flow when you can. Pull when you can’t.” In its simplest form, the Lean principal of pull says that no one upstream should produce a good or service until someone downstream asks for it.
In theory, this eliminates the muda of both overproduction and wait time. In practice, however, pull implies that the desired resource be delivered immediately over the value stream and arrive in the customer’s hands ready to consume out of the box.
But how can an IT department loaded with batch-oriented specialists possibly deliver something as complicated as a server instance at the very moment it is needed?
Software-defined infrastructure — the embodiment of lean pull
Software-defined infrastructure eliminates the rework and wait times by offering LOB managers a self-service way to get the IT resources they need right away. IT creates the IaaS resources one time, then publishes it as a plug-and-play blueprint within a strategic infrastructure software tool like StrataCloud SDI Install.
When an LOB manager needs to provision a new server instance, all he has to do is locate the preconfigured blueprint and click. A high-performance back end engine will perform all the steps the IT specialists would have taken and execute them. The new marketing cloud instance is ready in a matter of seconds without so much as a phone call to IT.
SDI in practice
Tribune Media is a high-profile software-defined success story. The company owns and operates television and digital news, entertainment and sports properties all over the web. In 2014, the company began replacing legacy infrastructure with new software-defined compute, network, and storage stacks with tremendous efficiency.
Tribune CIO, David Giambruno explains why he opted for software-defined data center for his operation. “…We designed the next-generation of enterprise IT with the goal of delivering a ‘frictionless’ enterprise—from infrastructure to applications,” he said in a recent interview with Baseline Magazine. “The SDDC enables what I call ‘indiscriminate computing’—the ability to move my stack and all of its data anywhere I need it.”
Flow is the preferred value chain delivery method for data-centric enterprises. But when flow isn’t possible, the Lean principle of pull delivers efficient customer value while eliminating waste that has long plagued IT departments in delivery of new cloud infrastructure.
By transferring the burden of intelligence from the hardware to the software, SDI has become the vehicle of choice for distributed enterprises to compete on Internet time.