Let the community know what you think. Share your opinions now!
The ability to ensure and enforce data quality.
Transparency and traceability via end to end data lineage.
Ability to automate acceptance and regression testing.
In my opinion, the following steps need to be considered before we move on to integrate systems,
* What systems that need integration.
* The limitations of the software
* How will the solution affect the employees?
* Which vendor will support which applications?
* How much will it cost to integrate the applications and what will be the impact?
* How quickly can I go to market?
Answer: Maximum flexibility within cost constraints.
Whilst a cloud data integration platform with open standards is preferable for maximum digital growth, if cost is the prohibitive factor then any short term integration should have future flexibility applied to it.
The Most important
1. Security (both cloud and connectivity to on prem)
2. If you are looking for volume data
a) performance (throughput)
3. availability %
4. frequency of updates
6. Enterprise connectivity support
As mentioned already 1st and foremost is data security whilst inflight and when at rest (cached). But not just technically but also the human processes around security such as data centre operations - think of SSAE16 for example.
Next up is execution location, depending on the data it maybe subject to data protection rules or contractual residency.
Once past the security/hosting/compliance/certification related questions, then need to think about capacity/throughput and latency. If your integration is part of a user experience you don't want the integration to flow through solutions quickly. That means you need rapid scaling to manage fluctuations in demand.
After that it is connectivity & ease of delivery.
One aspect that is important is evaluating what kind of data you are willing to put on the Cloud, security and performance are still issues that must be considered. For performance you need to consider how is the data going to be consumed (once a day, real time ?) and the volume that needs to be available on an ongoing basis. I always recommend companies to keep some current data On-Premise and new non-sensitive data on the Cloud and start growing that way, a hybrid system seems to work the best. Plus if the internet goes down, you can still run your business with local systems.
compliance and data security
How do each of these affect performance, and how does one decide which of these to use?