The Top 5 Data Architecture Mistakes To Avoid
The goal of the IT Architecture Journal is to help both IT architects and those who are considering hiring architects. Both of these groups of stakeholders must be aware of not only the fundamental concepts that underpin architectural practice, but also where and when architecture may go wrong. Today, we'll look at some of the more common challenges and faults that arise in the field of Data Architecture. While no clear legislation or even standard practices exist in any of the following areas, we have attempted to clarify where standard practice exists or is growing.
Issues of Context
A significant chunk of today's Data Architecture is entirely generic. It's not just a matter of whether the effort is high-level or not; even higher-level analysis and design might be targeted or contextual. Part of the difficulty is that many companies strive to standardize best practices based on industry-defined standards. This is very frequent in the DMBOK environment (Data Management Book of Knowledge).
There's nothing wrong with the DMBOK or the goal of standardizing data engineering solutions and design methodologies based on industry best practices right now. What's problematic is that many people take that information at face value without comprehending that it was designed to appeal to the widest possible audience. It, like every standard practice or technology guideline, is only as good as its relevance to any unique circumstance.
So, how valuable is just general information? As in other situations, the response is that it can be a terrific beginning point, but it should never be the final destination. XML and JSON, for example, are two different ways to create data messages, but knowing the underlying JSON and XML standards is only valuable if you know how to apply them to your own data (which standard to choose, and how it will be used, structured and integrated into your solution).
Too many Data Architecture initiatives give generic data with no "contextualization." This is a surefire way to have a project fail.
A Self-Contained Approach
When it comes to data engineering services, how do you know where to start and where to stop? This isn't merely an intriguing philosophical topic. When you think about it, data is nearly everywhere in the firm, but a lot of it - possibly the bulk of it - isn't handled as part of the "Data Architecture." Data from the following categories can be included:
- Content as well as email (including Social Media)
- Metadata
- Logs from the system (for any and all systems)
- Portfolio management or EA tools
- Smaller databases or technologies that aren't within IT's formal responsibility (often business users)
- Information about mobile and online apps
- With their own databases, applications (commercial software packages)
- Device data (and/or embedded databases)
You will almost certainly be asked to focus on only one'slice' of an organization's broader data environment as a data engineering solution providers. This gaffe, on the other hand, is more about the company's approach to data design and management as a whole. If a firm considers some of its data assets to be outside of its corporate data management processes, there is an issue. There is a problem when data management is further split between semi-autonomous groups within the organization. Both of these practices are common in today's society.
These circumstances lead to the establishment of data silos or islands. Some of the unavoidable effects are as follows:
- Duplication of data across the organization
- a deficiency in the ability to use truly effective analytics
- The expense of information technology is much higher.
The simplest way to knock down these walls and connect the data islands is to promise up front that the Data engineering solutions will encompass all data important to the firm so that design and management may be aligned with enterprise goals and expectations.
An Ineffective Governance System
It's important to note that we didn't just say 'Governance.' Everyone in the data engineering solutions industry loves to talk about governance, but what does it actually mean? Typically, governance consists of a workflow technique and a set of tools. If either the tool or the workflow process has limits or is non-existent, the Governance effort is likely to fail. "Actionable" refers to a collection of criteria that includes the following elements:
- All essential system management tasks must be connected with data governance.
- Security standards and methods are interwoven with data governance.
- The processes of data governance and portfolio management should be combined.
- The two concepts of data governance and data architecture must be combined.
- Adherence to carefully defined parameters that are regularly monitored. (ALM metrics, as well as integrity and performance measures, can be included.)
- There are clearly defined responsibilities, as well as authority and decision-making barriers, in the governance process.
- Governance is intertwined with both tactical and strategic organizational goals.
The Value of Data Architecture Is Underappreciated
This isn't just a Data Architecture issue, as you might anticipate. So, what precisely do we mean when we say architecture isn't taken seriously? Some examples of how this shows itself are as follows:
- No money has been spent on architecture.
- Limited or one-time investments (periodic reviews but no dedicated architecture practice)
- There are no plans to purchase architecture software.
- Other key procedures, such as Portfolio Management, are not integrated into the architecture.
- The architecture methodology was not used.
Like anything else, you get out of architecture what you put into it. If an activity's resources and dedication are constrained, the value derived is likely to be lesser than expected (unless it was understood to be a token effort from the start). In many ways, the Architecture lifecycle mimics the ALM (application lifecycle management) paradigm. Large-scale design that addresses all conceivable aspects of a network of individual solutions is referred to as architecture. Transferring Architecture into an Agile approach is tough, aside from time-boxing delivery requirements.
Architecture without method, on the other hand, is ineffective.
Concentrate solely On The Product
Here's where a word of warning comes in handy. If your project is entirely focused on deploying a particular product or technology, the product architecture is bound to take center stage. This, on the other hand, is a problem that affects the entire firm. An important benefit of an IT Architecture project is the ability to separate capability from the specific technology that will ultimately express that capability. What does this have to do with anything? Because hundreds of technologies will very definitely be involved in a vast environment.
- Because things and technology change.
- Because the Architect is more often than not a neutral 'change agent,' rather than a promoter or representative of a specific technology. The "Honest Broker" is another name for this occupation.
As a result, the most typical error in this field is believing that a single product or technology can provide a "Silver Bullet" solution that will substantially boost the company's performance. While this does occur on occasion, it is not typical. The genuine scope of a data engineering service providers is defined by the issue space associated with the organization's goals rather than the administration of particular commercial software packages in most circumstances.
Comments
Post a Comment