How the IT Department should develop a technical Roadmap

BKJIA, Internet headlines in February 9] Almost all IT departments must develop a set of targeted technical roadmap. After all, without the guidance of the road map, the purchase of new infrastructure hardware will inevitably lead to the choice of too high or too low configuration and combination, which will inevitably bring about high additional costs. But in fact, such guidance documents often do not exist, and the reason is very simple:

How can we establish a continuous technical plan to respond to all new situations in an ever-changing era?

The trouble in the whole process is that, in order to improve a set of planning, it is necessary to supplement and revise the infrastructure in three to five years so that it can fully fit with the theoretical solution; however, when the above work was completed successfully, we would suddenly find that the original plan was outdated. This inference alone is enough to make most IT departments choose to give up and focus most of their energy on the passive response measures with headaches and pains-but this solution also has problems: this will not only inevitably lead to excessive budget expansion, but also cause various difficulties in coping with the increasingly outdated hardware facilities.

In fact, you don't have to be in this situation. For a rapidly changing future, we not only have a way to plan, but also fully regard the work we have done as a new cornerstone. I will share with you the eight steps below, which have achieved remarkable results in practice.

Step 1: clearly define the planned Lifecycle

First, clearly define the time range to be covered by the technical roadmap. This decision-making process should be based on a number of factors, including how long the Organization will conduct a large-scale technological innovation and what requirements or necessary reference factors are included in the long-term budget plan ).

Generally, the cycle is defined.Should not be shorter than two yearsIf you cannot reach this level of foresight, the actual value of the plan itself will be difficult to meet your expectations. Similarly, it covers a period of time.Not more than five years-- In general, we cannot imagine where the technical situation or institutional needs will go after such a long period of time.

Step 2: Establish a set of macro infrastructure Questionnaire

Next, we need to define a questionnaire to clearly list the basic requirements in our infrastructure. In this step, you don't need to think too much about whether you can provide a perfect solution for each problem. The key is thatRaise Questions. In addition, you do not have to delve into some specific technical aspects. Over time, many answers will come to the fore.

On the contrary, we focus on higher-level issues. You may find that, although most of these problems seem to fall into the technical category, they are essentially irrelevant to technology. The following are some macro examples I have personally experienced:

• What is the year-on-year growth of data compared with last year?
• Compared with last year, what is the year-on-year growth in computing power?
• Do we know the new applications, solutions, and technologies that will be available soon?
• Do enterprises have any long-term expansion plans, such as deploying new offices )?
• What self-requirements do we have in terms of normal running time and reliability?
• What are our requirements for RTO/RPO, that is, the recovery time and recovery target point?
• After the configuration is completed, the configuration will be rich. What you need to do is to find answers to key questions that will guide us, even third parties, to design a new infrastructure that can bear the necessary load and meet the requirements of the company's own policies. If you do not provide all the information for achieving the above objectives, this list is still incomplete.

Step 3: seek answers

This is the most interesting part: the first attempt to enter our own questionnaire. Some of these problems, such as data growth rate, may need to borrow some ofMonitoring toolsTo answer. Others may need to have a frank conversation with the company's management, forcing them to share the actual figures in policies like RTO/RPO with us.

Step 4: design a future-oriented infrastructure

Once the questionnaire is developed, the next step is to design a new infrastructure that can meet the needs of various planning periods and must take the average annual growth rate into account. This step is also vague. Even areas that are already familiar with the past must be compared with new terminals as a necessary factor in the scheme.

When performing this step, make sure that no specific technology or product is involved, because these technologies are likely to be subject to large-scale changes within the timeframe of the road map. As you may have been accustomed to describing how we would need to deploy an EMC Clariion CX4-240 with 95 TB of storage and 30 thousand business processing capabilities per second, however, in this questionnaire, make sure that the module is described as"Entity Storage Solution", And clearly mark the required performance index. In terms of cost calculation, the current market price of the entire device is summarized and used as a reference data for the cost plan.

If you do not have sufficient bandwidth or practical experience to design your own work within the enterprise, you may wish to assign it to a trusted consultant or expert to complete the work. Since the reference materials at hand cover all the information in step 3 and Step 4, even a third party who is not familiar with the company's situation is fully capable of design.

Step 5: implement procurement

The next step is to submit the Facility scheme for the vision target to the enterprise management team, specifically, to the project owner ), check whether they are attracted by your long-term infrastructure planning and whether they can accept the overall cost. At this time, if they say they need to strictly implement the five higher-level reliability standards (9, 99.999%), it means that this solution is basically recognized. If they do not say so, we must continue to improve the planning and revise the expectations of the decision makers for the production effect, and minimize the cost of infrastructure construction as much as possible-of course, it is also possible to continue to argue, as long as they can make them realize that each step of the design can indeed bring benefits to enterprise operation, then the theory will become logical.

Step 6: perform gap analysis

After the implementation of experimental deployment, we need to accurately identify the gap between the existing infrastructure and the long-term infrastructure design. The essence of this step is to define the current infrastructure status quo to the Planning and Development Goals which need to be adjusted and changed-Please try to answer the following questions:

• How many additional virtual hosts do I need?
• How many vswitch ports do I need to deploy?
• How much disk capacity do I need to prepare for the new primary storage environment?
• Is the storage platform we are currently using capable of scaling to the target level?
• How much additional throughput/storage capacity is required for my data protection environment?
• Can the data protection environment we are currently using be expanded to the target level?

Step 7: establish a road map and prepare for implementation

After determining the gap between the new and old facilities, you can proceed with the implementation plan-and, of course, do not forget to put on the agenda how to gradually eliminate the gap during the implementation process. Now we can start with a budget item that meets the current price situation and select a specific product and specification for the facility.

Before we start to select a specific product, we will consider the long-term requirements in detail. This will allow us to examine many options from the top to the top, and select the objects that can best withstand the test of time. Taking the primary storage environment as an example: the Clariion CX4-120 will be hard to sustain as the business and operation volume continues to grow; as planned, in the end, we may need a device that is similar to the VNX5500 processing capability. In this case, we should not choose VNX5300 under the influence of short-term benefits, after all, it cannot provide us with excellent processing capability in a consistent manner. Furthermore, we can come up with the necessary information to prove the merits of this choice during discussions with the management, well-developed data reserves will make it easier for us to influence corporate decision-making.

Step 8: Maintain the Roadmap

It is often said that a good planning process will never end. Since we have already made so many painstaking efforts to build a road map, we should also get a corresponding return at this time-upgrading this complete road map will not bring too much workload. All you need to do is to regularly review the answers to the questionnaire and evaluate the changes in the current situation compared with the original one. If changes do exist, modify the final design as soon as possible to ensure that all changes are covered. Then, these changes will be incorporated into the infrastructure update work through the gap analysis mechanism.

As for the plan execution cycle setting, we need to review the frequency of the entire plan. I have seen some organizations review the plan once a month, and some organizations review the plan once a quarter, what's more, the review is performed only once every six months. Note that the longer the time interval between each review, the more difficult the review is to execute.

Comrades, bright future but tortuous path

As the sub-title says, it is never easy to develop a plan for infrastructure development. The key lies in the fact that every second we spend on making a plan will bring tangible results and benefits. Although this specific solution cannot meet everyone's requirements, some of these steps may have broad guiding significance. When the target condition changes rapidly, the implementation scheme starting from the end of the plan and ending at the beginning of the plan can effectively avoid Rework in the deployment process.

No matter what you deploy or how you deploy it, do not leave the technical plan completely behind. Even if it only helps us avoid a planning mistake, such as the selection of SAN alternatives, the entire design process is worth the money.

Original article: 8 steps to building and maintaining an infrastructure road map

Bkjia.com is an exclusive special article. For more information, see the original author and source .]