Skip to main content

Hello Everyone 😊,

 

I'm now working on an endeavour that will integrate our organization's multiple old systems with Cognite Data Fusion (CDF). Although there are numerous advantages to CDF's current architecture, there are certain obstacles we must overcome in order to guarantee smooth data transfer between CDF & our more antiquated systems—especially those that weren't created with contemporary data platform in mind.

I would appreciate your thoughts on the following specific queries and worries:

Data Connectivity: How can CDF and older technologies that use antiquated communication formats or protocols create and sustain reliable connections with one other?🤔 Have anyone of you run across problems with particular kinds of systems?🤔 If so, tell us about it.

Data Consistency and Quality: When importing data into CDF from older systems, how can you maintain and guarantee data consistency?🤔 Exist any particular Cognite ecosystem tools or procedures that support preserving data accuracy and consistency?🤔

Real-time Data Integration: Enabling real-time data connectivity between CDF & our older systems is one of our objectives. What are the most effective ways to accomplish this, and how can you deal with any delay problems that may occur from the outdated technology in legacy systems?🤔

Security Considerations: Security is a big concern because of how old some of our systems are. How can you be certain that vulnerabilities won't be revealed while combining these systems with CDF?🤔 Do you advocate any particular CDF security setups or protocols?🤔

 

Any guidance, insights, or resources you could provide on these difficulties would be highly appreciated. Our goal is to create a solid and future-proof integration, thus it would be really beneficial to gain knowledge from those who have addressed servicenow related problems.

 

Thank you in advance for your help and support.

On connectivity:

With CDF you have access to a collection of Extractors that pull data from external systems into CDF. Some of these are antiquated, though they are mostly modern systems. When connecting to a system without explicit support you generally have a few options:

  • Connect through some common protocol. We support ODBC, OPC-UA and OPC, many sources are in some way covered by these natively.
  • Use a third party tool to connect, such as PTCs KEPserverEX, these let you bridge the gap between the source system and our tools.
  • Create your own integration using our tooling, see Integrate > Extractors > Create your own extractor in Fusion.

On Real-time:

Typically we batch data for about a second, but are otherwise realtime. If the source itself is delayed, there is nothing you can do short of going to a different source for the data. Not sure I fully understand this question.

On Security:

Usually extractors will run in the “DMZ” part of the Purdue model for securty, we carefully distinguish between extractors that only read data, and connectors that may write data back to the source as well.

Beyond that we always advocate for using encrypted traffic and solid authentication.


Reply