Blog

Tokenization: Selection Criteria

By Adrian Lane

To wrap up our Understanding and Selecting a Tokenization Solution series, we now focus on the selection criteria. If you are looking at tokenization we can assume you want to reduce the exposure of sensitive data while saving some money by reducing security requirements across your IT operation. While we don’t want to oversimplify the complexity of tokenization, the selection process itself is fairly straightforward. Ultimately there are just a handful of questions you need to address: Does this meet my business requirements? Is it better to use an in-house application or choose a service provider? Which applications need token services, and how hard will they be to set up?

For some of you the selection process is super easy. If you are a small firm dealing with PCI compliance, choose an outsourced token service through your payment processor. It’s likely they already offer the service, and if not they will soon. And the systems you use will probably be easy to match up with external services, especially since you had to buy from the service provider – at least something compatible and approved for their infrastructure. Most small firms simply do not possess the resources and expertise in-house to set up, secure, and manage a token server. Even with the expertise available, choosing a vendor-supplied option is cheaper and removes most of the liability from your end.

Using a service from your payment processor is actually a great option for any company that already fully outsources payment systems to its processor, although this tends to be less common for larger organizations.

The rest of you have some work to do. Here is our recommended process:

  1. Determine Business Requirements: The single biggest consideration is the business problem to resolve. The appropriateness of a solution is predicated on its ability to address your security or compliance requirements. Today this is generally PCI compliance, so fortunately most tokenization servers are designed with PCI in mind. For other data such as medical information, Social Security Numbers, and other forms of PII, there is more variation in vendor support.
  2. Map and Fingerprint Your Systems: Identify the systems that store sensitive data – including platform, database, and application configurations – and assess which contain data that needs to be replaced with tokens.
  3. Determine Application/System Requirements: Now that you know which platforms you need to support, it’s time to determine your specific integration requirements. This is mostly about your database platform, what languages your application is written in, how you authenticate users, and how distributed your application and data centers are.
  4. Define Token Requirements: Look at how data is used by your application and determine whether single use or multi-use tokens are preferred or required? Can the tokens be formatted to meet the business use defined above? If clear-text access is required in a distributed environment, are encrypted format-preserving tokens suitable?
  5. Evaluate Options: At this point you should know your business requirements, understand your particular system and application integration requirements, and have a grasp of your token requirements. This is enough to start evaluating the different options on the market, including services vs. in-house deployment.

It’s all fairly straightforward, and the important part is to determine your business requirements ahead of time, rather than allowing a vendor to steer you toward their particular technology. Since you will be making changes to applications and databases it only makes sense to have a good understanding of your integration requirements before letting the first salesperson in the door.

There are a number of additional secondary considerations for token server selection.

  • Authentication: How will the token server integrate with your identity and access management systems? This is a consideration for external token services as well, but especially important for in-house token databases, as the real PAN data is present. You need to carefully control which users can make token requests and which can request clear text credit card or other information. Make sure your access control systems will integrate with your selection.
  • Security of the Token Server: What features and functions does the token server offer for encryption of its data store, monitoring transactions, securing communications, and request verification. On the other hand, what security functions does the vendor assume you will provide?
  • Scalability: How can you grow the token service with demand?
  • Key Management: Are the encryption and key management services embedded within the token server, or do they depend on external key management services? For tokens based upon encryption of sensitive data, examine how keys are used and managed.
  • Performance: In payment processing speed has a direct impact on customer and merchant satisfaction. Does the token server offer sufficient performance for responding to new token requests? Does it handle expected and unlikely-but-possible peak loads?
  • Failover: Payment processing applications are intolerant of token server outages. In-house token server failover capabilities require careful review, as do service provider SLAs – be sure to dig into anything you don’t understand. If your organization cannot tolerate downtime, ensure that the service or system you choose accommodates your requirements.
No Related Posts
Comments

Good read Adrian.

Token Generation Mechanism is another important criteria for selection of a tokenization solution. Is there enough randomization and seeding in the tokens so their generation criteria can not be predicted. Can the tokens be generated as FPT (Format Preserving Tokens); can they have a 1:1 relationship for data integrity within an enterprise; can the 1:1 relationship be broken in certain situations like tokenization of salary information; can they be generated for multiple data assets; can they be made to pass or fail Luhn checks to minimize application modifications; can they be visually distinguishable; can they have site designators for different regional locations; can they have designators to prevent token collisions when tokens are being generated across multiple data stores or vaults: just a few token generation criteria which we continuously strive to cover as we proliferate tokenization within our growing customer base.

Abir Thakurta
http://blog.nubridges.com/

By Abir Thakurta


If you like to leave comments, and aren’t a spammer, register for the site and email us at info@securosis.com and we’ll turn off moderation for your account.