Backlogs and Constraints in Continuous IT Operations

Continuous Improvement in an Omni-speed IT estate  needs IT Operation Managers to re-look at what can be done different, effectively. IT Operations has borrowed time tested methodologies from manufacturing operations, and adapted them. In this article we will focus into the science behind bottleneck and constraint relates to IT Operations.

Imagine an operation of filling a bucket of water. You go to the tap, and It takes 5 minutes of continuous flow to fills it. If there is dirt in the tap filter, blocking the rate of flow, you would need more time to fill the same bucket. If you go to a tank and put the bucket in and pull out a bucket full of water, the same task could be done in seconds. What does it tell you?

Let’s dig in deeper in the same scenario.

Scenario 1: Your task is to fill water in a bucket. When you go to the tap, the tap has a limitation of how much water it can flow and thus it needs 5 minutes to fill a bucket. That is its constraint. If we need to fill the bucket in a second, we go to the tank. We could achieve that by alleviating the constraint, skipping the tap and going to the tank. You have a choice, but you cannot change the constraint.

Scenario 2: Your task is to fill the bucket and you go to the tap. Due to dirt in the filter, the bucket fills slowly and needs more time than the usual 5 minutes to fill. This delay is a bottleneck, holding up all the other subsequent task. To still be able to meet the time, one has to clean the dirt stuck in the filter and regain the desired flow of water. We could achieve this by taking the action to remove the cause of the bottleneck. You have control over the backlog.

This analogy can be applied to a day to day operations, in IT or Manufacturing, to explain why some process builds inertia and become inefficient.

The bottleneck theory helps to identify problems and create solutions for streamlined operations. Do note, when a process outpaces even by a single step in the overall processing, it causes bottleneck. This is generally a supply and capacity driven problem. Just like in manufacturing, IT also would need continuous monitoring to look out for backlog accumulation causing bottleneck.

The Theory Of Constraint

One general tendency is to increase capacity to address bottleneck. This would soon create a situation of surplus if not calibrated for the right speed. Speed for an IT process would not be constant, and thus for an Omni-speed IT operation, we will need to design for the flexibility in service. Where you as an operations manager get to increase or reduce services on demand.

How does a working process gather inertia?      To answer this let us look at the Supply. Say you have process to service requests. The process may have multiple process handles that could be manual intervention, autonomous handling, workflow etc. Imagine, due to an event, we have increased numbers of requests, say a project go live. One of the situation could be that the existing process handlers are stretched. Another scenario could be that within the capacity to serve was reduced, say due to a natural calamity causing outages. This will add constraints to serve. In both cases we will see accumulation at the slowest process handler in the IT operations. This will result in process gaining inertia and not perform to its design. Changing constraints over time adds to inertia. 

Some easy quick fixes could be to identify the bottleneck and adding additional capacity to process handle. This may elevate the current bottleneck, but can introduce newer ones. This is where we need to do a flow time analysis of a process. Flow time is defined as the amount of time a flow unit spends in a business process from beginning to end, also known as the total processing time. If there is more than one path through the process, the flow time is equivalent to the length of the longest path. But flow rate is an average rate, and not the peak rate. It will help baseline your existing process design, when you work on its improvement.

How do Theory of Constraint help re look at an IT process and help its redesign?      This is a methodology for identifying the most important limiting factor (i.e. constraint) that stands in the way of achieving a goal and then systematically improving that constraint until it is no longer the limiting factor. There is a continuous cycle involving the five steps.

• Step 1:  Identify the constraints in a process. At this stage you identify the goal, desired throughput, look out for cause, identify your inventory of tools and process.
• Step 2:  Decide how to exploit the systems constraints. Identify the methods to be used and decide how we maximize the throughput. The method should be evaluated to meet the goal of the improvement.
• Step 3:  Subordinate everything else to the decisions of Step 2. Bring focus of all available resources, tools and processes on remediation of the constraint.
• Step 4:  Elevate the system's constraints. Here bring in the changes, investments and process improvement. At this stage a decision to invest by management is taken on better tools, resource training, revamped process etc,.
• Step 5:  Evaluate and if the current constraint is broken, go to step 1. Study whether solving the current constraints created other constraints. Do not allow inertia to set in. The process has to be monitored carefully as to whether other constraints now exist and to monitor the progress of the old constraint.

How do we identify constraints and bottleneck in operations?      Constraint specifically refers to a factor outside of the operation manager's control. A machine working at full capacity represents a manufacturing constraint and similarly a process handle working in full capacity is its constraint. Even an employee shortage can be a constraint. Bottleneck, on the other hand, is used in operations to refer to something that is temporary in nature. With a few smart adjustments, bottlenecks can be eliminated. And if they cannot be eliminated, they are actually constraints.

The approach to evaluate an IT process optimization will need a study into the details using Theory of Constraints to identify constraints, Study Bottleneck to identify accumulations, Flow Time Analysis to determine optimal paths, time & motion study to study process handle to seek more efficient methods of execution and then run simulations. Running successful simulations runs are necessary as any problem alleviation in a certain process handle do not assure that a new problem will not appear elsewhere. Thus when we do simulations, we should be able to not only address current problems, but also identify and fix potential problems.