One of the main benefits of these shows is that you get the time to wander about and meet interesting people you might miss otherwise. For example, there’s an interesting use case of the ONE Platform right opposite the SPARKL booth, where the Austrian firm NTS are showing their Quality-of-Service (QoS) definition tool.

This thing works by letting the user create a network policy across a bunch of routers - for example, a policy to prioritise traffic from one node to another. You can then apply that policy - which automatically propagates it across a bunch of routers so it takes immediate effect - and unapply it thus releasing any resources it may tie up.

This uses the OnePK API on supported routers to achieve its effect, but it’s basically manual through a UI. We had a great discussion about how with the SPARKL Transaction Router we can tie these useful, high-level policies to transactions - or even individual events within transactions. We went so far as to design the markup we’ll put into the transaction config to make this happen.

So the Transaction Router makes these important QoS policies completely automatic - in the context of any transaction that’s executing. When the transaction finishes, or alternatively when the associated service is brought down, the QoS policy is also automatically torn down.

This is really nice, because it means we can say to a customer that certain policies will be applied automatically when the appropriate business-level transactions occur. It’s QoS with automatic lifecycle. Furthermore, we can point to the audit logs generated by SPARKL which provide the evidence that the policy was applied correctly - should the customer need to confirm that at some future date.