Delphix

Delphix keeps hold of a single, continuously updated copy of each of your production databases that you can provision virtual copies against as required. You can refresh, reset, or rewind your personal copies at any time, and you can also bookmark, branch and share them. Notably, provisioning is incremental, meaning that when your production database is updated, only those updates (and not entire datasets) will be pushed out to your virtual copies during a refresh. Even then, refreshes are strictly on demand: your users are given the freedom to decide if and when they need an update. In addition, multiple databases can be provisioned in a synchronised fashion. This could be used, for example, to facilitate integration testing for apps that use data from multiple data sources. What’s more, the platform’s “data time machine” functionality provides a complete record of all production deltas, meaning that at any time you can look back and see exactly what changed, and when, in your production data.

If your data sources contain sensitive data, you will also need to profile and mask it. In Delphix, this functionality is generally applied when you first create a copy of any given production database, creating a masked version of it that desensitised virtual copies can subsequently be provisioned from. Moreover, profiling and masking in Delphix can be leveraged on structured and unstructured data alike, can be scheduled or run on-demand, and come equipped with preconfigured templates tailored to some common use cases, including applications (such as SAP) and regulations (such as GDPR). It also supports tokenisation, audit and compliance reporting, and stand-alone masking (without data virtualisation). Finally, if you already have a preferred data masking solution, you may use that instead.

Delphix also provides the Data Control Tower, a single location for controlling your data automation efforts across your Delphix deployment and enterprise-wide. Automation in this case is API-driven: Delphix even describes its platform as “API-first”.

Data virtualisation provides several advantages over other test data management approaches, most notably that it allows you to leverage a complete (and compliant) copy of your real production data, without worrying about representative subsets and the like. It also lends itself very well to efficient provisioning and operationalisation of test data, and from there to DevOps, reinforced by the product integrations and self-service capabilities provided by Delphix.

In terms of functionality, we would rate the product’s masking and profiling capabilities as capable, but not exceptional, although being primarily policy driven, we acknowledge that they can be applied effectively to a large quantity of data and data sources. Moreover, the data virtualisation, data masking, and data profiling capabilities Delphix offers are all automated, highly scalable, and provided as part of a single platform, making Delphix a one-stop-shop for virtualisation-driven test data management.

Delphix is also interesting for the fact that it offers specific, point functionality that is well suited to test data management, but at the same time does not limit itself to that space. Indeed, although it must be said that accelerating application development is the company’s most common use case, we are told that they are increasingly in demand as part of broader CI/CD toolchains. Given this, the product’s heightened emphasis on integration makes a lot of sense, with the addition of comprehensive APIs in particular allowing it to be leveraged remotely and/or via application code. The upshot of all this being that you can use Delphix to provision test data wherever in your toolchain that you happen to need it: a powerful capability.

The Bottom Line

Delphix is no longer the only vendor that offers generalised test data management predicated primarily on the merits of data virtualisation. In this sense, it has lost the advantage of exclusivity. That said, it is still the most mature and well-known solution of its kind, and it almost certainly has the most experience working with large enterprises. This gives it an appeal all its own.

Delphix holds a single, continuously updated copy of your production databases (including all binaries, configuration files and so forth). It can then provision complete virtual copies of this data as required. The automation and self-service that this provides fits well as part of Agile and DevOps initiatives. Features include the ability to refresh, reset and rewind data, as well as the ability to bookmark, branch and share data. Notably, Delphix provisioning is incremental, meaning that when your production database is updated, only those updates (and not entire datasets) will be pushed out to your virtual copies during a refresh. Even then, refreshes are strictly on demand: your users are given the freedom to decide if and when they need an update. In addition, different databases can be provisioned in a synchronised fashion. This could be used, for example, to facilitate integration testing for apps that use data from multiple data sources.

Fig 2 Self-service in Delphix

If your test data is potentially sensitive, you will need to a) profile your data and discover which tables, columns and so on are, in fact, sensitive, and b) mask that data. How this works is that you take your copy of the production database, profile it, create a masked, virtualised version, and then provision virtual copies from that version. This can all be achieved from within the Delphix platform. In addition, both data profiling and data masking via the Delphix platform can be leveraged on structured and unstructured data, can be scheduled or run on-demand, and come equipped with preconfigured templates specific to particular applications (such as SAP) and regulations (such as GDPR). If you already have your own data masking solution, you may choose to use that instead. Another option is to use internally written scripts for masking, though this is not an approach that we would recommend. Stand-alone (without data virtualisation) masking support is also available for mainframe (iSeries, VSAM and z/ OS) systems. Delphix also supports a tokenisation capability, and audit and compliance reporting is provided.

Data virtualisation is entirely unlike other approaches to test data management. It does not seek to produce a scaled-down set of test data that is in some way representative of your production data, whether created via data subsetting or via synthetic data generation. Rather, data virtualisation enables you to distribute – and therefore test on – all of your (virtualised and masked) production data. There is clear value in this. Beyond the fact that it ensures that your test data is 100% representative of your production data, it also lends itself very well to effective distribution and operationalisation of test data, and from there to DevOps. The self-service capabilities that Delphix provides reinforce the latter point.

Fig 3 Masking sensitive data in Delphix

Moreover, the data virtualisation, data masking, and data profiling capabilities provided by the Delphix platform are rapid, automated, and highly scalable. In terms of functionality, we would rate the product’s data masking and profiling capabilities as capable, but not exceptional. However, they are primarily policy driven, and hence can be applied effectively to a large quantity of data and data sources. In addition, all of these capabilities – along with self-service – are provided within a single, unified platform. This makes Delphix as a whole easy and convenient to use for test data management, as all of the tools a user might want are located in a single place.

The Bottom Line

Delphix is one of a very small number of data virtualisation vendors that operate in the test data management space. Although there are several reasons to recommend it, ultimately your opinion of it will be determined by whether or not you like its data virtualisation centric approach to test data management. As we said in our last report, “if you like the approach, you’ll love the product.”