Argent provides a number of different levels of security. Argent’s products are used by the U.S. federal government agency called Intelink, which links the CIA, FBI, NSA, and other agencies they don’t want to mention…

All communication across the TCP/IP links is encrypted. Argent philosophy to encryption is to provide an architecture that can be updated as security improvements are made – when 256-bit encryption is replaced with 512-bit encryption, so Argent’s design enables the new encryption components to be added in under 30 minutes.

On a more prosaic basis, Argent’s scheduling products provides multiple levels of access, so different levels of access can be granted, from read-only access for a sub-set of jobs, all the way to SuperUser or Admin rights. Argent’s access levels reflect the real world where different people require different rights.

2 – Stability

Without a running product, all is lost. Argent ensure stability in two ways: design and scalability testing. From a design viewpoint, stability is addressed by planning how Customers will use the product.

Customers’ need vary widely – one Customer simply needed a reliable scheduler with full auditing to run three jobs per day at 4:01 p.m. A production job scheduler of the power of Argent may seem to be massive overkill, until you learn these three jobs were SWIFT transfers to the New York Fed (http://en.wikipedia.org/wiki/SWIFT). For each hour these transfers were late the fine was $100,000. Opps. No room for error here. Thus Argent.

At the other extreme is a company that is a service bureau for credit unions; this firm acts as the credit unions’ data center. With very strict SLA requirements, this service bureau run — on a typical day – 8,000 jobs per day, and at the busy end-of-quarter over 25,000 jobs in a day.

With Argent’s central design, stability is an integral part of the product.

3 – Scalability

Argent’s basic tenet for scalability is encapsulated in Argent’s Rule of 100. The Rule of 100 simply means Argent products are tested against 100 times the expected maximum load – if the maximum load is expected to be 25,000 jobs in a day, then Argent attempts to run 2,500,000 jobs. This approach really shakes a product down, and it also highlights the weaknesses in debugging (or lack thereof).

Argent’s internal record stands at a little under 3,100,000 jobs in a day running on 24 real machines (that’s 129,000 jobs per Argent Queue Engine per day, or 1.49 jobs per second per Argent Queue Engine). Needless to say, the street lights in the neighbor dim and flicker, and performance is not lightening fast, but Argent handles this gigantic load.