High-Performance Resilience

Monterey delivers resilience at two levels: high-availability and disaster avoidance. Both have unique characteristics that differentiate Monterey from alternatives.

High Availability with No Impact on Throughput

All high-availability solutions depend on duplication in order to make a component redundant, and ensuring high-availability at the application level is no different.

Conventional high-availability solutions implement duplication in a way that imposes trade-offs between throughput, integrity, and costs.

For example, in order to ensure zero data-loss, conventional duplication must be synchronous, which can dramatically affect throughput (especially where it takes place over a wide-area network). Alternatively you can relax data integrity to minimize the effect on throughput by making duplication asynchronous, but now you risk data loss, which you may have to mitigate through expensive enhancements to your infrastructure. These problems are exacerbated if you need more than a single duplicate in order to achieve the required level of resilience.

By contrast Monterey uses its overlay network to dynamically insert as many additional instances of the redundant component as desired into the traffic flow. This exploits the overlay network’s uni-directional flow to guarantee that all duplicates are updated without affecting throughput. So, for example, adding a single duplicate for redundancy purposes merely requires a one-way network hop to the duplicate before the duplicate invokes the next step in the processing chain.

Adding an additional duplicate (i.e. primary + 2 fall-backs) merely extends the daisy-chain of duplicates by one. Again there is no impact on throughput – even if a wide-area network is involved – because at no point is there any blocking or waiting for a response (i.e. achieves the benefits of asynchronous replication but with none of the disadvantages).

Whenever a primary or fall-back component fails, Monterey’s policy-driven management framework will automatically:

• re-wire the overlay network to bypass the failed component;
• re-create the failed component;
• dynamically re-incorporate the re-started component into the overlay network.

…at no time is there any interruption or pause to service while this takes place.

Consequently you are now able to provide high-availability with zero data loss, with higher service levels and at a fraction of the cost of alternative approaches.

Disaster Avoidance

Complementing the unique high-availability features offered by Monterey is its ability to proactively avoid disaster. A combination of Monterey’s policy-driven controls and elastic location management means that impending failures can be identified and the application components in harm’s way can be very quickly, and automatically, re‑located without interrupting service.

This is also true for network bottlenecks and failures.

Consequently a higher level of service can be maintained for your application users at lower management costs.