Roland Bullivant is accustomed to the stunned silence that often greets a demonstration of Safyr®, Silwood Technology’s Metadata Discovery software. Moments after realizing that Safyr creates a complete Metadata Glossary for large ERP and CRM packages in a matter of hours, a process that can take an entire team working full time six months – or more – many potential customers are a bit emotional. One customer said simply, “I’m grieving for the lost years.”
Safyr is not an Enterprise Metadata Management (EMM) tool. Rather, it’s “Metadata Discovery software,” said Bullivant, Sales and Marketing Director at Silwood Technology. He and Nick Porter, Silwood’s Technical Director, talked with DATAVERSITY® about how their clients are using Safyr to ‘discover’ Metadata and condense months of work into minutes.
Companies using Enterprise Management solutions like SAP, Oracle, Salesforce, or Microsoft, face months and months of work to create a Data Dictionary. Yet Bullivant said Safyr can create:
“A complete Metadata Glossary for SAP systems, [that is] searchable, readable, usable, shareable, and accurate, based on what [a client has] actually implemented in less than three hours from connecting to the system. You can’t do that manually. You can’t even do it in three months manually, regardless of how many people you throw at it.”
A Brief History
Bullivant said Safyr evolved out of a need for Metadata Management from companies using erwin Data Modeler with large customized ERP and CRM packages, like SAP or Oracle.
“Our first real exposure to Data Modeling was with erwin way back in the day, and then over a period of time, we discovered that we were selling solutions around the erwin Data Modeling tool and helping organizations use it effectively, and what we discovered time and time again was that whilst erwin is really good at reverse engineering database schemas, handling SAP and other large packages needed a different approach.”
Bullivant said there are a number of reasons for this. “Basically, there’s no Metadata of any value in the database itself.” Also, the Data Model in an SAP system is about 90,000 tables, and even if there was valuable metadata, “you probably don’t want to reverse-engineer engineer 90,000 tables as it will take time and resources.”
The highly customized nature of larger systems was also an issue. Companies they worked with were spending “huge amounts” of money and time trying to understand the data structures underpinning SAP applications, for example. Safyr was developed in 2002 to solve that problem, so that companies with large CRM and ERP systems can deliver information, Data Management projects, and Metadata Management quickly, more effectively, and with greater accuracy.
“Since then, we have acquired a network of partners all the way from IBM through to organizations such as ASG, Datum, Adaptive, and many others” he said. erwin also resells Safyr, and Silwood also has technical partnerships with other organizations. They also work with SAP specialists, and others who also sell the product. These partnerships are its primary go to market vehicle.
The company’s customer list includes BP, Boeing, Shell, World Bank, RS Components, Hydro Tasmania, Department of Homeland Security, Tesoro, and Volkswagen. “We only sell to people who have got these big packages and the big packages tend to be bought by big complex organizations,” They need to understand the tables, attributes, relationships and definitions underlying their data systems in order to maximize their return on the data they hold and deliver new programs. he said. “So that’s what we do. We solve that problem.”
What is Metadata Discovery?
Bullivant stressed that because Safyr is in its own unique category as an ‘information provider,’ there is an incorrect assumption that it does ETL or Data Lineage. He provided an example to illustrate the unique gap Safyr fills within existing software in the marketplace:
“ASG have a product called Data Intelligence which used to be known as Rochade, and the product could scan all sorts of files and databases and systems, and structured data – and I think unstructured data – to bring that Metadata into their repository, to do Data Lineage and Semantics, or whatever they want to do with it. Well, one thing that the product didn’t do is know how to understand, or reverse engineer PeopleSoft, Salesforce, SAP, or anything like that, and we do that. So that’s our job.”
Organizations faced with General Data Protection Regulation (GDPR) compliance, for example, must be able to locate and know how to manage the personal data of any EU citizen on demand, anywhere it appears in order to comply, or face steep fines. With implementation of an Enterprise Metadata Management program, a Master Data project, or even a Data Warehouse project, he said, “The question that is going to come up at some point is, ‘Where is the data? Where do I go to find that data in the first place?’” For smaller systems, “it’s relatively straightforward,” and it’s possible to use a tool and reverse engineer it easily, “but for those big packages, it’s a whole order of magnitude more difficult, and that’s the problem we solve,” he said.
Porter adds, “But we’re not an integration layer. We are an understanding technology, so it’s for discovering the location of data in these packages.”
According to Porter, it’s a double-edged sword, being the only product of its kind. “We did work with Gartner there for a while and they were unsure where to put us, partly because we’re the only ones that do this.” While this should provide an advantage in a marketing sense, Safyr is still a somewhat well-kept secret:
“If you don’t know that there’s a product – or a category of product – out there that can give practical insight into the data structure of these packages, you won’t go looking for it. So, when your systems integrator says, ‘Oh well, we’ll have to spend six months mapping out your Data Modeling for your SAP system,’ if you don’t think there is software that would do that, you’ll be unlikely to argue with them, right?”
There are tools like Oracle’s Enterprise Metadata Management solution, which Porter says, according to his understanding, will ingest Metadata from all sorts of places, with the exception of its own packages. “The last time I looked – and it may change – but I believe SAP is the same. They don’t really have anything that would give a data professional the ability to access and use Metadata from SAP with Information Steward, for example, in an easy way.”
With so many tables, it’s an impossible task even if you knew what you were looking for, he said. Many other products can connect to SAP but, Porter said:
“It’s sort of like landing in the middle of New York and someone has taken all the street signs away. You wouldn’t know how to get where you are going, because you wouldn’t know what all the street names were.”
Signs that say ‘One’ and ‘Two,’ but without ‘East,’ or ‘West,’ don’t provide enough information, he said, “So you could be wandering around for ages, because just being able to connect to something doesn’t necessarily mean that you know what it is you are supposed to connect to.”
Porter said that customized systems further complicate the process of using acquired information:
“Everybody connects to SAP and everybody connects to Salesforce. Everybody connects to databases and to the packages. But, unless you understand the Data Model and the structures underneath that, it’s really difficult to navigate and figure out what you want and where it is, unless you happen to know someone who knows, or unless you are lucky and you can search the internet for a Data Model that actually reflects the same version and customization that you’ve got, or unless you are lucky enough to find someone who can sell you a template which has the exact version and customizing.”
Many of Silwood’s clients are currently dealing with GDPR compliance. Attempting to find personal data in an enterprise system without a clear understanding of the Metadata is quite difficult, said Porter. Using Safyr, “If you want to find ‘birth’ in an SAP system, everywhere that ‘birth’ exists as an attribute in all the tables in SAP system – we can find it. It takes us three minutes.” He said it’s often difficult for prospects and new customers to understand how that’s possible:
“People often imagine that we are a kind of a pipe that plugs into each of these packages at one end, and then at the other end, plugs into erwin, or ER/Studio, or PowerDesigner, or ASG’s Data Intelligence. Whilst strictly that’s correct, part of the problem there is that while we enable understanding of these big packages, what you don’t want to do is move 90,000 tables and three quarters of a million attributes into those environments because all you’re doing is moving the problem from one place to another. A lot of the functionality of Safyr is about intelligently scoping the Metadata from these products before you move it somewhere else.”
For example, to scope a bill of materials tables from an SAP system and then to utilize those in erwin or ASG, it becomes a two-part problem, he said: How to identify the business data associated with those concepts, and then how to scale a 90,000-table Data Model into a manageable size. A Data Model of that size, while it might work:
“Is going to be completely unmanageable, and humans are not good at handling Metadata of that size and scope,” he said. “What you need to be able to do is reduce it down into business artifacts and business concepts.”
Porter says there are other products in the marketplace that move Metadata from one place to another, or can give a limited view of SAP metadata. “Safyr is more than that,” he said. They don’t see products that move Metadata as competition:
“A large part of the capability of our product is about doing that intelligent scoping, working out the relevant set of tables, attributes, relationships, and business definitions which are appropriate to a particular business artifact, so yes, we are not in the Enterprise Metadata Management space in terms of having a general software for managing Metadata. We are a vital part of any project that needs to do Enterprise Metadata Management where it applies to one of the package types that we address, like SAP or PeopleSoft.”
Bullivant showed Safyr to a project architect a couple of years ago from a very large consulting firm. About 15 minutes into the demo, the customer went very quiet, he said:
“I asked if I had lost him, if I had gone wrong somewhere, and if he wasn’t following, but he said, ‘No, the product is great, the product is fine. My problem is that I have ten people full-time working on a project at a SAP customer for at least the next six months, doing what your product does in an afternoon.’”
Bullivant understands, because he’s seen it before.
Photo Credit: Sergey Nivens/Shutterstock.com