I can guarantee your favorite web sites use memcached and/or Redis as part of their back-end infrastructure. Both technologies are popular ways to store data that needs to be easily accessible, but both require compromises (as any technology does). GarantiaData, a Tel Aviv-based startup, wants to eliminate some of those compromises, namely by perfecting a way to compress the data stored in both Redis and memcached.
The startup, which was a GigaOM Structure LaunchPad finalist, has built a service hosted on Amazon’s Web Services (s amzn) that provides a way to scale out a memcached or a Redis data store “infinitely” without adding latency. With this promise it has raised $3 million in funding from European angels since its founding in March 2011.
Ofer Bengal, the co-founder and CEO of GarantiaData, explained to me that the company’s software is better able to compress the unstructured data that companies store in memcached or Redis, which then makes it possible to run those as a service using Amazon EC2. The software also manages the deployment of new EC2 instances so the data store can grow as needed based on the customer demand. For memcached, Bengal says the software can compress the data up to 90 percent as opposed to the roughly 10 percent compression most people can deliver on memcached servers. Compression of data isn’t offered on Redis, although people do compress their data before they put it in Redis.
The compression means that Garantia can charge per gigabyte for storage and still make money by actually compressing and storing less data, while it also gives customers access to Amazon’s platform. And if that platform is big enough for Netflix, (s nflx) it’s probably elastic enough for most. The service just launched its public beta, and prior to that launch Bengal said the service had about 25 customers storing up to 10 Gigabytes each. That pales in comparison to the terabytes stored in Facebook’s memcached servers, but this is a service more likely to appeal to a startup that wants to scale out on an as needed basis while it tests the waters for its app.