distribute access to internal APIs. Different vendors ship distinct building blocks and integrations. They can be SaaS or be installed in your infrastructure.
server code • Distinct API views and responses from the same origin • Call composition between APIs • Single point of access • Leverage migration from legacy code • Leverage breaking monolith to microservices
routes • Manage connection pools and integration into legacy systems • Batch call stored procedures • Integration to message bus and messaging solutions • Serverless architectures (e.g. AWS Lambda)
are needed for your deployment • Pick and choose based on budget and in house knowledge • Migrate from an existing reverse proxy based on features • Hosted or SaaS: evaluate the team workload and devops skill set • ** Features may change - conduct your analysis based on features you need
or in the face of a vendor change, dependency on BaaS might have created space so other development tasks were taken by the team. Take care to not outgrow your backlog counting on a long term contract. Alocate BaaS dependency as technical debt. Authentication model Engage into market proven authentication models that might be present in other players. Most APIs stick to OAUTH2 or AUTH Secrets.
have competitive price models but make sure that you can serve and front APIs outside of their infrastructure for an attractive network ingress/egress price. Development model The configuration and coding should be versionable outside of the tool and automatable - meaning you should have tests for your configuration changes. The Gateway should support dev, integration and production environments and version promotion.
instrument your backend. If you host your Hosted model on cloud: make sure you reserve the minimum or regular day to day usage and pick marketplace prices to lower the impact on your monthly rates Hosted on VPS/bare metal: plan for at least 30% peaks on each server. Backend savings API gateway should save backend computing and network resources. Make sure you review your sizing and scaling. Do use automation to ensure resources grow according to the workload.
reducing their footprint on your architecture 2. Use standard transport and authentication protocols 3. Take care with cache and data coherency 4. Strive to be able to migrate to at least one open source solution