Most of my recent work has centered around one common theme – model based solutioning; as we look at the new drivers Service Integrators must understand before proposing a solution the steps we have taken start to slowly make sense.
There is nothing really new here; it’s common sense.
we do this everywhere….
so, as we look to create new solutions (solutions as code), we can connect every point in these methods
So what have I done so far?
Architecture was key and from a number of older projects Archimate became THE notion standard for us; tried and failed attempts at using (new) OTS products failed to scale, but the meta model stayed true. Then came Neo4j and Graph….
A graph database provides the means to connect and understand information in new and unconstrained ways; differing from traditional databases through a NoSQL structure, Graph databases are able to make not only 100’s but 1000’s of relationships between essentially line items, these relationships are further enhanced by exploring the near neighbour of nodes which can unlock the hidden information within the overall dataset. Figure 1 below shows the basic structure of graph database.
Figure 1: Basic Graph data structure
The current DXC knowledge graph solution contains all the information required for our client engagements; through industry and technology trends we are able to discuss and develop conversations with our clients which are relevant to their business, the planned business value roadmap module will build upon this information and outline a business strategy for our clients. The final element of solutions not only provides the proof points of our ability to deliver, but also opens the door to a new way to create solutions; through solution composition. By utilising the building blocks documented in existing solutions or creating new building blocks during design and development, we can compose new leading edge solutions with known levels of assurance through rapid and agile methods. Figure 2 below shows the overlaps of these 3 areas.
Figure 2: DXC knowledge graph clusters
The knowledge within this platform requires buy-in from across the organisation; but the return is also across the full organisation, as shown in figure 3 below.
Figure 3: Knowledge contributors and benefactors
Overview of current solution
The DXC knowledge graph has been developed to capture and exploit 3 areas instrumental in the creation of leading edge and emerging solutions; with a 4th module planned to create our client roadmaps.
- Industry and technology trends
- Delivered solutions
These 3 areas are presented across 3 web modules and 2 supporting Microsoft HoloLens Applications.
Digital Explorer presents the industry and technology trends and Solution Explorer allows users to submit, discover and view solutions, whilst Graph Explorer allows users to navigate across both knowledge sets via a graphical user interface. HoloLens also provides a means view the Digital Explorer and Solution Explorer content within augmented reality.
The EMEA XTech program provided a new and enlightening knowledge set to help shift the client conversation away from legacy IT services and towards the line of business. By tracking industry trends and disruptors against a standard timeline (see table 1 below), we not only present our own viewpoint within each industry but can also use this information to gain an understanding of where the client is on their digital journey. This is further enriched within the Graph database as a single trend can have a relationship to one or more industries (see figures 4 & 5 for industry and technology models) ; thus, allow us to compare and select solutions from other industries. Supporting the industry trends are a set of technology trends mapped against the same timeline, allowing users to bridge between business, technology and solutions.
|Lifecycle Stage||Description||Adoption levels|
|Emerging||Idea inception, market research and feasibility investigations||0-5%|
|Leading Edge||Adopted by challengers with no legacy systems or processes||5%-10%|
|Early Adopters||Ambitious companies who are quick to follow||10%-25%|
|Mainstream||Widespread adoption across leading players in the market||25%-50%|
|Late Adopters||Catch up of the market to near universal adoption||50%+|
Table 1: Trend lifecycle stages
Figure 4: Business trend Graph metamodel
Figure 5: Technology trends Graph metamodel
Simple web interface
A common layout to represent the trends within each industry; as shown in figure 6 below.
Figure 6: Example of industry web page
Allows industry and technology experts the capability to contribute, discuss and share trends and disruptors.
Pan industry disruptor viewpoint
The connected trend and disruptor knowledge provides the means to understand the maturity of similar trends across multiple industries. This can be visualized directly within the graph explorer as shown in figure 7.
Figure 7: Example of a pan industry view against a business disruptor (taken from Graph Explorer)
Open Data model
Knowledge is available for anyone to consume via a set of API’s; this has been used within the initial HoloLens application to present all industry trends (figure 8)
Figure 8: View of the industry knowledge through Microsoft HoloLens
Solution are our primary asset, representing our actual deliverables to our clients, in essence our solutions are already compositions of many building blocks from our practices and partners. What We Build\Solution Explorer has outlined a simple reference model (see figures 9 & 10 ) to capture and present any solution we deliver for our clients, from demos to full production solutions. The model focuses on the business motivations and enabling features within the solutions, allowing other users to review and validate the potential of the solution for their clients. Through events such as Solution Olympics we have been able to gain new insights into our existing solution, such as common patterns and shared features.
Figure 9: Solution Metamodel Model (Archimate)
Figure 10: Solution Metamodel Model as represented within Graph
Ability Promote key solutions and internal programs
The initial landing page for Solution Explorer ensures both first time and regular users are aware of the latest content available, for example the finalists from Solution Olympics are presented on the landing page, as shown in figure 11 below.
Figure 11: Solution Explorer landing page
Common data model
The solution meta model has been designed to support the means to capture the motivations and features from any solution, within any industry.
Similarity based search
The Solution Explorer search engine breaks down the model into the 5 core areas; the general solution description, value related objects, feature related objects, industry and finally all other fields and applies the users search term against these 5 areas; this is further enhanced by the usage of Wordnet to provide similarity matching for words within the search
Dynamic architecture diagrams
Each solution is presented as a dynamic diagram (see figure 12) within a solution datasheet, allowing users to understand the features; their type (category) and relationships. Each diagram can be filtered based on any solutions profiles defined within the solution.
Figure 12: Example solution diagram with Solution Explorer
Alignment with internal taxonomy
All common areas of the solution model such as industry, practice and categories are aligned with the master data management records within the organisation.
Industry and practice viewpoints
The graph database and explorer modules allow users to gain insight into pan industry and practice collaboration and solution development.
Open Data model
Solution models are available for anyone to consume via a set of API’s, this has been implemented through a supporting BOT using Microsoft Skype, as shown in figure 13 below.
Figure 13: Functional view of the skype bot using the graph solution API’s
One key advantage of the graph based solution is the ability to understand where our key resources are having the biggest impact. Stepping away from Opt-in knowledge models and user profiles, through direct contribution we are able to understand automatically those who are trend spotters, trend developers, leading and emerging solution developers. Further extended through the connections and neighbour relationship within Graph able to expand this across industries, practices and solution categories as shown in Figure 14 below.
Figure 14: Person view within Graph
Current content and release timeline
Figure 15 below illustrates the progression to date across these 3 area within the current solution.
Figure 15:Progress to date for the solution model and knowledge graph
The potential for this graph based platform is wide reaching, extensions to the existing modules and the introduction of two new modules has been outlined as shown in figure 16 below.
Figure 16: Development roadmap for Graph and progression towards Composer
Key new features
Digital and Solution Explorer
Sharing our knowledge
The web applications presenting this knowledge are good, however the ability to generate a controlled document on demand based on this content is key. The solution and industry PDF generator will allow our client teams to produce up to date content for our clients to review offline (this will include the appropriate document framing)
Industry guided tours
Extension of the current site guided tour; tailored to each industry, we can provide a walkthrough of key industry trends as defined by our industry teams, concluding with example solutions from Solution Explorer, this will clearly show how we not only understand the industry but have a proven track record of delivering.
Feature suggestion bot
Building upon the trend suggestion bot, the ability to suggest and reuse existing features from other solutions is incredibility powerful; creating not only a viewpoint of where and how our own offerings have been used, but gain before unknown awareness of potential emerging offerings or patterns.
The reuse of a feature would generate a duplicated node within the graph model, ensuring any future changes to the child or parent feature does not have a negative effect through the inheritance. This would be tracked through the “specialized” relationship as defined within Archimate as shown in figure 17 below.
Figure 17: resulting data view of the feature suggestion bot
Business drivers and trend suggestion bot
The relationship across the two primary knowledge sets can be further extended through directly suggesting a business trend based on the client drivers captured during solution harvesting. As defined within Archimate, a business driver is “defined as something that creates, motivates, and fuels the change in an organization”. This is a direct match to the industry trend captured within the business trends dataset and therefore can support the solution owner to identify the best matching business driver during the creation of their solution entry, the created relationship is shown in figure 18.
Figure 18: Resulting data view of the suggestion bot for matching client drivers to industry trends
Through the combination of reused features and data analytics we can identify common patterns from across the full solution dataset. These patterns can in turn be extracted and created as super-features within the graph model. Figure 19 below, illustrates an example from the existing solution dataset
Figure 19: View of identified patterns or common features across multiple solutions
These super features can provide not only extended supporting information for the engineering and deliver teams, but also a foundation set of capabilities for the creation of new solutions.
Archimate Export & Import
The underlying solution model is aligned with the Archimate specification; the ability to both export and import solution models will further extend the potential reuse and value to our build and delivery teams.
The graph platform and its content provides the foundation towards even greater insights and capabilities for our clients, figure 20 outlines the next steps within the development roadmap.
Figure 20: New modules roadmap
Business Value Roadmaps – Target Mid 2017
The current industry knowledge is our internal viewpoint, by using this same knowledge to create our client roadmaps, we transform this to higher value information; not only enhancing our viewpoint but providing key information to our CT, Build and Sell organisations. The early identification of emerging demands can help direct future investments to the correct areas, whilst also allowing us to gain earlier insight into areas with diminishing demand; thus, giving us the opportunity to invest in people, tools and processes to meet future market demands. Figure 21 illustrates how a single trend can be tracked and viewed across multiple client roadmaps, multiple this information across 200+ accounts and the value and potential only increases.
Figure 21: Graph view of confirmed client demand/interest against a single business trend
The client roadmaps will be developed within an intuitive user interface as shown in figures 22 and 23 below.
Figure 22: illustration of a user developing a client roadmap in a customer workshop
Figure 23: Developing a client roadmap via a tablet
Solution Composition is the ultimate goal of the modelled based work; upon building a library of solutions, patterns, features, business trends and drivers this can be transformed into applied knowledge for the creation of new solutions; the “Assured solution development through a model based approach” invention paper outlines a graphical user interface to capture, analyse and suggest matching elements from within the graph database, figure 24 provides a view of one potential composition canvas; paired with knowledge of the current environment detailed work orders can be generated to engineering team. Work orders which detail which common feature to use, how to use and integrate and extend the solution to meet the client requirements.
Figure 24: mock-up of composer solution canvas
- Composer Concept Video: https://www.youtube.com/watch?v=xqo8ozz5iIA
The metamodel under pinning the repository is aligned to the OpenGroup Archimate 2.1 specification and standard and WordNet  provides the data dictionary used for the similarity based search.
- Project Tiger – Model Once, Use Many: HPE IP Fusion ID: 83861998 (Stevens, Flower; 2014)
- Tiger Canvas – An intuitive solution to create, manage and peer review proposed business solutions: HPE IP Fusion ID: 700217432 (Stevens, Flower; 2014)
- Project Tiger Maturity Model for Managed Service Providers: HPE IP Fusion ID: 83817891 (Stevens, Brown 2014)
- Service Model Domains for Service Integrators: HPE IP Fusion ID: 700216908 (Stevens, Rogers, Knight, O’Brien; 2015)
- Assured solution development through a model based approach: HPE IP Fusion ID: 700216908 (Stevens, Rogers, Gill; 2016)
- Solution Fit – A method to asset the potential of a business solution based on keyword search: HPE IP Fusion ID: 710225008 (Stevens, Keane, Piorun, Giebultowicz)