Building a Tagging System and Filter Pattern

Following the launch of FDC 1.0, I engaged in various exploration projects.

During the strategic planning for the product roadmap, a highly sought-after feature emerged: the incorporation of functionality allowing the selection of multiple servers based on similar criteria and subsequently adding them to a Tenant or during environment composition.

Recognizing the significance of this feature, I dedicated efforts to conceptualize and integrate it seamlessly into the evolving roadmap, aligning with the product’s continuous enhancement and user requirements.

In the Fungible data center (FDC), a tenant’s internal teams can access servers and deploy workloads of cloud computing by selecting and combining groups of servers.

With an excessive inventory of devices in data centers nowadays, to be competitive in the market cloud service providers (CSP) are geared up to meet the increasing need for an easy and efficient way to manage all kinds of devices (mostly servers in the FDC settings). Flavors come in handy when defining the compute, memory, and storage capacity of fungible computing instances.

In Fungible data center, there are two types of metadata:

• System-generated metadata are automatically generated from server properties.
• Custom metadata are managed by users.

I used a concept map as a visual tool to organize and structure the functionality and user flows within the system. This map served as a guide to understanding the interrelationships between different components and features, ensuring a holistic perspective during the design process.

Understanding Users

Considering Fungible’s business model, our users will include data center managers and company IT managers, and the pattern will be used across various roles and features.

Here are some assumptions I made to predict user behavior:

1. Users likely have some familiarity with filters and tags, so we need a flow with common patterns.
2. Users need a simple filter for quick access to their daily tasks.
3. Users should have some flexibility to create their filters and even use programmable queries.
4. Users might want to save searches for regular reuse.

Define User Tasks

I created the product brief based on the user scenarios:

1. Manage (add, edit, or delete) custom metadata on servers
2. Filter servers using multiple criteria, such as AND or OR (boolean), multiple values, and different operators.
   1. Data center managers need to filter servers with the same criteria to manage and maintain devices.
   2. Data center managers need to filter and select servers to allocate to logical partition.
   3. Company IT managers need to filter and choose servers to add to an environment with a group of homogeneous instances.

To start, I did a competitive analysis of filter patterns in multiple products.

Inspired by the research, I came up with 3 options, each based on the flexibility and ease of use.

Option 1

Option 2

Option 3

Finalizing Design

After discussing the options with the product and engineering teams, I concluded the design for the initial phase and outlined development plans for the release of Fungible FDC 2.0.

Project Roadmap

After initial dev hand-off, I also worked with product and engineering to figure out a development plan for the tagging and filtering feature.

Phase 1: Implement metadata management and simple filter, build design system pattern.

Phase 2: Implement advanced filter using query.

Phase 3: Apply pattern to GPU.

• Embrace an iterative design process, recognizing that features will evolve and may shift over time. Plan ahead for potential functionality add-ons to accommodate future needs. Design with scalability and flexibility in mind to ensure adaptability to changing requirements.
• Consider the impact of design decisions on development timelines and plan accordingly for efficient implementation. Regular communication between product, design, and engineering teams helped a lot to maintain alignment on goals and priorities throughout the project.