TimeXtender

PAGE 16
PANDORA MAINTAINS
ITS GLOBAL MARKET
WITH CONTROLLED DATA
OPERATION PROVIDED
BY TIMEXTENDER’S
HOLISTIC DATA SUITE

IRI
PAGE 18
PRODUCE DATA
PRODUCTS WITH

MODERN DATA
VERACITY WITH
VORACITY

MANAGEMENT:
Succeeding with
Data Mesh &
Data Fabric

Best Practices Series

 Achieving Success
With Data Mesh
& Data Fabric

DATA MESH AND DATA FABRIC ARE KEY TECHNOLOGIES and frame-
works that are advancing efforts to achieve data democracy. The Legacy modernization to cloud-based and web-
obstacles to data democracy are plentiful, however. Effectively scal- friendly systems has been ongoing for years, but with
ing data and analytics to ever-shifting business needs is one such the move to data analytics and AI accelerating, the
challenge. Another is being able to deliver services and capabilities
need for modernization has never been so urgent.
through increasingly complex and connected systems, while assur-
ing security. Adding to this challenge are data silos that cut off
end users and critical systems from the data needed to deliver design concept,” the DBTA study notes. “Data quality platforms
predictions and the insights needed to run digital business, both are being focused on improving operational processes and
today and tomorrow. managing data quality rules, though integration challenges
Above all, there is a push to deliver data democracy in order are significant. Moreover, becoming proficient with data man-
to leverage the power of analytics at all levels of the organiza- agement tools is necessary.”
tion. To advance opportunities in data democracy, two promi- A similar pattern of adoption—and hurdles—is seen for data
nent approaches to decentralized architectures and frameworks mesh, emerging in a survey of 200 data managers published in
are emerging—data mesh and data fabric. Data mesh offers a September 2024 by BARC (Business Application Research Center).
decentralized, domain-oriented framework, while data fabric “Despite the clear value proposition, widespread implementation is
supports a unified, centralized architecture that provides seam- progressing slowly,” the BARC study’s authors concur. “Compared
less access and integration from all points in the enterprise. to last year, the share of respondents stating that data mesh is
More than 1 in 4 enterprises, 26%, recognize the benefits of ‘relevant and being done’ only increased slightly to 20 percent.
data fabric for IT data management, while 6% see its advantages However, the percentage of those planning to implement data mesh
for business analytics and discovery, according to the “2024 Mod- has dropped significantly from 35 percent to 24 percent, while the
ern Data Management & Data Fabric” market study produced by group considering it of little or no relevance has increased by 12
Unisphere Research and Radiant Advisors in March 2024. percentage points to 27 percent.”
Yet the implementation of data fabric faces hurdles, such as When looking at data mesh, which offers significant potential,
“transitioning from legacy systems and addressing data gover- particularly in terms of improving business agility and satisfac-
nance,” the DBTA survey shows. tion, the BARC study “shows that implementation is still slow
“Key components such as data virtualization and metadata in practice.”
activation are identified as essential for effective data fabric imple- The bottom line is that data managers and analysts “are skep-
mentation, underlining a cautiously optimistic view towards this tical about data mesh in an above-average proportion of cases: 37

14 BIG D ATA Q UA RT E R LY | WINTER 2024

 on silos and may introduce latency in efforts to make data
widely available across enterprises. In addition, moving away
from traditional data architectures and frameworks to more
decentralized models needs to be planned out. “Organizations
should plan strategic migrations considering technological
and human factors, ensuring smooth and well-supported
transitions to the data fabric,” the DBTA report urges. “Invest-
ment in training for new tools and technologies, especially
around advanced features like knowledge graphs, is crucial to
fully leveraging the potential of the data fabric.”
• Strive for open, advanced architectures. As the DBTA survey
documented, “Noteworthy percentages for real-time unified
platforms, data fabric and data mesh architectures suggest
a shift towards more complex, interconnected systems that
facilitate data accessibility and interoperability across different
business units and applications.”
• Educate the business on the benefits of and differences
between data mesh and data fabric. Data mesh and data
fabric represent a departure from the standard treatment of
data through databases and data warehouses. About 20% of
respondents to the DBTA survey, for example, either dismiss
data fabric as “just vendor marketing or hype” or simply do
not understand it. The survey report states, “As understanding
Best Practices Series improves… there’s a noticeable increase in positive percep-
tions about the potential of data fabric for IT data manage-
ment and business analytics.”
percent consider it of little or no relevance.” At the root of this is • Adopt a “data product” approach to data management and
overall “resistance to decentralization,” the BARC authors found. delivery. Data products are trusted, reusable, and consumable
“This is despite many organizations facing data and analytics data assets that are managed by individual business owners
challenges that data mesh could effectively address.” and are intended to be accessible to end users across the enter-
To overcome potential resistance and challenges to developing prise. “In essence, data mesh envisions the future of data and
data mesh and data fabric capabilities, managers should consider analytics through decentralized data ownership, where busi-
the following measures: ness domains are responsible for creating valuable, desirable
• Rethink the value of data in the organization. It’s recog- and feasible data products,” according to the BARC authors.
nized that enterprises in today’s economy require timely and “This approach is grounded in a self-serve data platform and
accurate data from a variety of sources to compete. This calls federated computational governance, which balances central
for ramping up investments in data infrastructure. Orga- policies with decentralized responsibilities.”
nizations need to “understand and strategically invest in a Ultimately, the BARC survey shows that 73% of enterprises
broad spectrum of technologies that support robust, scalable actively seek a decentralized data architecture that can be deliv-
and intelligent data ecosystems,” the DBTA report asserts. ered via data mesh and data fabric. At the same time, much of the
“This approach ensures that the data fabric supports current hesitation toward adopting data mesh and data fabric is not just a
operational needs and drives innovation and competitive technical challenge, “it is an organizational one that demands cul-
advantage in an increasingly data-driven world.” tural change and strong commitment,” the study’s authors state.
For both data mesh and data fabric, this means investing in “However, this commitment, particularly toward decentral-
cutting-edge technologies, as well as adopting new ways of izing data ownership, is crucial to becoming a truly data-driven
thinking about data. New technologies and solutions include enterprise,” the BARC researchers added. “Traditional D&A
ETL tools and data virtualization, metadata repositories and [data and analytics] approaches have not proven effective in
data catalogs, data governance platforms, data orchestration achieving this goal. … Decentralizing data ownership inevitably
and workflow automation, knowledge graphs, semantic increases the complexity of governing and managing data prod-
layers, and observability and monitoring tools. ucts and the underlying platforms. The solution to handle this
• Step up the move away from legacy architectures and sys- complexity lies in federated computational governance, which
tems. Legacy modernization to cloud-based and web-friendly strikes the right balance between central and decentral aspects
systems has been ongoing for years, but with the move to data in your governance approach, preventing chaos and ensuring
analytics and AI accelerating, the need for modernization effective management.”
has never been so urgent. Older systems are naturally built —Joe McKendrick

D BTA.COM/BIG DATA QUARTERLY 15

 sponsored content

PANDORA maintains its global market

with controlled data operation provided
by TimeXtender’s Holistic Data Suite
Danish jewellery company PANDORA offers its THE COLLECTIBLE, THE CONTEMPORARY
customers fashionable, hand-finished pieces In 1982, Danish goldsmith Per Enevoldsen and his wife
Winnie opened a small, local jewellery store in Copenhagen,
at affordable prices. But to help maintain its often travelling to Thailand for inspiration and materials.
rapid growth from a single Danish store to As they gained a reputation for sleek and contemporary
offerings, their focus began to shift from small-scale to
over 2,400 concept stores globally, it required
wholesale and the PANDORA brand was born. Over the last
a controllable way to access and analyse the 30 years, it has made a significant mark, creating high-quality,
increasingly large volumes of company data. collectible pieces that customers can afford. The company has
rapidly expanded, now operating in more than 100 countries
across six continents. What was once a single store has now
become the world’s best-known jewellery brand.
CUSTOMER BACKGROUND
» Danish designer, manufacturer and distributor of FROM DESIGN TO DISPLAY
contemporary jewelery Peter Hansen is a manager and global BI architect at
» Products sold in more than 100 countries across PANDORA. He is primarily responsible for making sure
six continents that the company remains compliant in the outlined data
» Known for their collectible charm bracelet architecture principles that support the data coming in from
company sources. “PANDORA does everything from design
CHALLENGES and innovation, to prototyping, manufacturing, distribution
» To support extensive company data collected and retail,” he says. “We buy up the raw silver and gold and
from manufacturing, distributing and retail turn it into jewellery.” Yet managing operations from start to
departments sparkling finish means tracking every stage, which results in a
» To maintain a functional and controllable lot of internal data.
data warehouse that can handle many individual As if that’s not hard enough, PANDORA was also
contributors and contractors finding it increasingly difficult to cohesively manage data
» To allow for less development and avoiding from numerous other diverse data sources. “We have five
dependencies different ERP systems that we pull financial information
from,” Peter explains. Additionally, PANDORA utilises
SOLUTION more than 20 different providers of retail information.
» TimeXtender was implemented to provide secure While everyone could see the immense value of having
and reliable management of diverse data sources accurate, up-to-date inventory information, since this
» PANDORA contractors are able to work with directly affects what jewellery pieces are available in-store,
many projects without disrupting other data the complexity of the data reports required for a clear
schemes overview of company operations was daunting. PANDORA
» Having access to TimeXtender’s operational data required a system that was both controllable and universal.
exchange to work with financial, retail, sales and One that would give them more time to finesse and
inventory information customise data reports, rather than bogging them down
with intricate coding requirements.
BENEFITS
» Fully automated and controlled documentation SAVING TIME WITH INSTANT DATA ACCESS
» No need to produce or write code manually With a large network of consultants and contributors,
» Quick start up allowing contractors to utilise PANDORA wanted a solution that would allow its 20 or so
platform functionality internal and external BI developers to be up-and-running
» Stable systems and reports supporting easy in a few hours and spend more time on modelling and
contractor collaboration analytics and less on coding. TimeXtender, being uniform
and automated, met these requirements. After implementing

16 BIG DATA Q UA RT E R LY | WINTER 2024

 sponsored content

TimeXtender, Peter noted an immediate reduction of time- “We save the hours not on the analysis
consuming data preparation duties. “We save the hours but on the operations to actually just
not on the analysis but on the operations to actually just
make sure we get the data. I’d say that we spend maybe make sure we get the data. I’d say that
only 20% of the time we’d previously spent on preparation we spend maybe only 20% of the time
before TimeXtender.” Peterhas also been impressed by how we’d previously spent on preparation
quickly the employees picked up on the functionality of the before TimeXtender.”
TimeXtender platform. “We’ve seen new contractors starting
with it in the morning of their first day and, after lunch, start Peter Hansen
being productive in TimeXtender.” Manager and Global BI Architect, PANDORA

A CONTROLLABLE AND STREAMLINED SOLUTION

PANDORA put a great importance on controlled access frequent updates. They want to know what’s happening
to data when looking to acquire quality information from down to each hour, or even quarter hour!”
multiple sources. With their manufacturing facilities in
Thailand, along with departmental offices in Germany, GAIN A 360º VIEW OF THE CUSTOMER
Denmark, Hong Kong and the USA, and with BI For PANDORA’s marketing department, being able to
development in all five locations, the company needed to combine information from ecommerce and point-of-sale as
secure the data being collected while ensuring that specific well as in-store systems that automatically count customers
employees had access to the data they needed. Peter was is critically important. TimeXtender has been utilised to
pleased to find that with TimeXtender, data was “provided combine consumer data with sales data, so that PANDORA
much faster to the people who need it.” He adds, “The can better answer the question, ‘Who actually buys our
configuration part is also easy, meaning that you more or products?’.
less have a drag-and-drop solution. And it’s very easy to
set up data flows, while data can be fully or incrementally A CONTINUOUS PARTNERSHIP
loaded.” This ability to create unique data flows catering to With the implementation of TimeXtender, PANDORA has
departmental needs is an enormous benefit, as Peter can been able to simplify data preparation, obtain key consumer
easily administer data flows and access controls for external insights and support their extended network of skilled
contractors. “You have a control interface with user access contractors. For Peter, it is the ease, efficiency and speed of
control, and you can easily set up security rights too. This TimeXtender’s data architecture platform that has become
allows us to utilise new resources relatively quickly.” instrumental in the company’s data analytics. n

EASE IN COLLABORATION
A company the size of PANDORA requires an incredible YOU’RE IN GOOD COMPANY
degree of collaboration among departments and external TimeXtender inspires quality business decisions
contractors. Here are Peter’s three practical examples of how with data, mind and heart. We do this for one simple
PANDORA uses TimeXtender to help manage projects: reason: because time matters. TimeXtender is a
recognized global software company enabling
• Creating independence—“The developers we use instant access to any type of data in the organization.
are mainly contractors, so of course we wanted a tool
that was controlled by us and not creating a dependency To learn more about TimeXtender and how the
on them.” power of automation can move your business
• Universal and controllable—“The fact that you don’t forward, visit: www.timextender.com
have to write much code in this means that the solutions
produced are very similar.”
• Easy access to current data—“In manufacturing as well TIMEXTENDER
as in our distribution centers, there is a demand for www.timextender.com

dbta.com/Big Data Quarterly 17

 sponsored content

Produce Data Products

with Veracity with Voracity
THE IRI VORACITY data management platform has been data search, profiling, classification, masking and cleansing
featured and described in multiple best practice and big data- (quality), plus basic cataloging and lineage. A number of data
related issues of DBTA Magazine. You may remember it for observability and access control capabilities are also included.
its ability to combine data discovery, integration, migration, Voracity Data Infrastructure platform features include:
governance, and analytics. Because Voracity maps data between support for polyglot source and target data stores, pipeline
multiple source and target formats and silos, it’s a platform you orchestration, CI/CD integration, data product versioning
may wish to consider using at the center of your data mesh. (via Git), PII and access credential encryption, and big
The data mesh is essentially a more recent paradigmatic data transformation performance that scales vertically
manifestation of other Voracity-supported data integration or horizontally. A number of graphical self-service data
architectures. Voracity can be used in ways not too dissimilar engineering and democratization features are also available
from a Logical Data Warehouse or Production Analytic through the DataSwitch web front-end.
Platform. What distinguishes a data mesh primarily are its
characteristics of: 1) Domain-driven ownership/design, 2)
Data as a product, 3) Self-service data infrastructure, and 4)
Federated computational governance.
As described in previous DBTA best practices issues around
Data Integration and Governance, Voracity can serve as a Data
Infrastructure and Federated Governance platform which
supports domain-driven design. If you review our Big Data
Packaging page, you’ll see that the core idea of Voracity as a
data mesh platform is the same—to integrate and govern data
from disparate sources to make informational output, or data
products, fit for data science use (and re-use).
In a data mesh, you wind up with something akin to data
marts, and in Voracity, you can—without a data warehouse
team—produce immediate data products or reports, too.
Integration, cleansing, masking, reporting are all same-
job, same-pass affairs. You can quickly wrangle your own
data products that contain trustworthy, business-formatted Beyond the functionality Voracity embeds, another key
subsets (or formatted reports) with clean, anonymous, and consideration is how it addresses a common problem with data
de-duplicated data inside. meshes: a lack of clarity around data ownership. When different
people are looking at different data products to find answers to
the same quantitative question, they will get different answers,
and usually after data has been replicated too many times.
A data mesh designed in Voracity can clarify data
ownership through metadata files which record the authorities
that engineering and business teams can ask about: common
data definitions, data discrepancies, inter-domain integration
points, and reducing unnecessary data normalization steps.
The aforementioned FAIR metadata in Voracity can also
help align data producers and consumers and standardize data
formatting. Data sources and accesses are registered in the
IRI Workbench design IDE and readily identified in job scripts.
Voracity follows “A Data Mesh Approach to Data Voracity data-product projects with those artifacts can be shared,
Warehousing,” as described in this 2021 Alexis Mackenzie compared, and discussed from a common repository like Git.
article in Towards Data Science—from FAIR principles This framework in Voracity is flexible and easy enough to
supported by SortCL data definition language files cataloging create small test projects quickly. By conducting a variety of
structured source metadata to using non-competing, real-time POCs, your company can find the best ways to guild governed,
read replicas supported by the IRI Ripcurrent module. efficient data products.
Voracity delivers several Federated Governance platform
features recognized in data mesh architecture, including: IRI VORACITY www.iri.com/products/voracity

18 BIG DATA Q UA RT E R LY | WINTER 2023