Chip memory, sectors and access keys

Serial-number-only is often enough

If your system stores everything centrally and only needs to know which band it is looking at, a chip returning a unique fixed number does the whole job. The band holds no data, so a lost band leaks nothing, and there is no memory to corrupt. Most zone-access and cashless implementations at resorts work exactly this way, with the entitlement held on your server rather than on the wrist.

Writable memory means the band carries state

Where a system needs the credential itself to hold information — an entitlement that must work when the network is down, a stored count, a room number written at check-in — the chip needs writable memory. That introduces the questions that come with any distributed state: what happens when a write fails halfway, how the band is resynchronised, and what a band says if it was written by a system that has since changed its mind.

Sectors and keys

Capable chips divide memory into sections that can each be locked behind their own key, so a lock system can hold one section and a payment system another without either being able to read the other. This is genuinely useful in a property running several systems on one credential. It also means key management becomes your problem: keys must be held securely, shared with integrators carefully, and changed when someone with access leaves. A property that cannot say who holds its keys does not have a secure system.

Do not over-specify

The most capable chip costs several times a simple one and, on many properties, is used exclusively to hold a serial number that any cheap chip would have returned anyway. Ask what your access platform actually reads. If the answer is the serial number, the expensive chip is buying nothing except the option to change your mind later — which may be worth it, but should be a deliberate purchase rather than a default.