Akamai Brings Key Value Data to the Edge, Adds API Acceleration
Content delivery network (CDN) services provider Akamai has expanded its edge serverless compute platform, EdgeWorkers, with the addition of data at the edge with this week’s launch of its EdgeKV key-value store.
Released into beta earlier this year, EdgeKV is now generally available and gives developers access to a lightweight database at the edge, rather than in data centers. In this most recent update to its product, Akamai has also launched API Acceleration, which pairs specialized hardware with optimized API routing and GraphQL caching to increase reliability and performance at scale.
CJ Arnesen, senior product manager for edge computing at Akamai, said that EdgeKV complements Akamai’s edge compute, something the company had offered previously in the form of pre-determined functions, but more recently updated to a serverless model with EdgeWorkers, which allows users to run JavaScript at the edge.
“It’s great to be able to run JavaScript at the edge, you can execute some code, but all that has to run in memory, it has to run self-contained,” Arnesen said. “We needed to add a data store to this so that customers can either write or read data, right there at the edge, reducing that latency to go to origin. It’s great to have edge compute, but part of that story is accessing data, or even writing data. This is allowing us to do more heavy lifting on the edge.”
A key-value store is a schema-less database that is designed to store unstructured or semi-structured data. There is no complex query language, rather, a user looks up a key, and they are returned a value. Values can be anything from a simple string to a nested JSON to a base64-encoded binary, but its simplicity offers reduced latency compared to relational databases. Arnesen explained that the addition of a key-value store to the edge means that companies can offer their users more personalized experiences, faster.
“The digital experience these days is robust. Brands want to give each user the best experience possible, a personalized experience, to hook them and get them engaged,” he said. “EdgeKV can store basic values about a user or about a region or about anything you want, and then EdgeWorkers can call upon that data and, in real-time, rewrite the webpage, do dynamic content assembly, and customize the page for that user. This really brings the power of origin-based personalization compute to the edge, reducing all the penalties you used to get with personalization.”
Arnesen further explained that EdgeKV could be thought of as “a global write and a local read,” with Akamai offering a global synchronization of data with a service level agreement (SLA) of 10 seconds. “The data that you need is always at your fingertips locally on your edge server, but we continue to synchronize and write it globally,” he said.
In addition to EdgeKV, Akamai also launched API Acceleration, which Alex Balford, a senior product marketing manager at Akamai, described as a response to a recent 30% year-over-year increase in API requests that brought Akamai’s traffic to more than 300 trillion API requests in 2020.
“APIs have been around for a long time, but as organizations have continued their digital transformation, they’re shifting from monolithic to microservices,” Balford said. “They’re deploying microservices in their cloud infrastructure, they’re also deploying more at the edge. All these are communicating over APIs. We have a lot of major customers that are utilizing APIs and need great performance and high reliability for that traffic.”
While API requests might be small transactions, Balford explained that they are often highly compute-intensive — doing things like setting up and tearing down Transport Layer Security sessions in a short amount of time. As such, they can benefit from the special-purpose hardware, reserved capacity, and prioritized routing that Akamai is dedicating to them with its API Acceleration.
Moving forward, Balford said that Akamai would be looking to not only help customers with API discoverability but also to integrate Akamai’s API solution with its security offerings. As for EdgeKV, Arnesen said that Akamai would be looking to augment its edge offering with the addition of a “more robust, or more traditional, database at the edge, as well.”
