To Boldly Go where No Cache has Gone Before
Enterprises deploy ScaleArc to provide continuous availability for their apps – we get a lot of kudos for delivering database failover that prevents interrupting the app. But a lot more than failover goes into making the database highly available, and performance is a big contributor. ScaleArc boosts database performance with many a variety of features. Today, we’re very excited to announce a major leap forward in our NoSQL caching technology, extending the types of data our customers can safely cache. We’re also sharing independent test results of the performance gains this new caching supports.
Today’s big advancement comes in how we let you invalidate cached data. To date, the ScaleArc software has let you have cache expire based on time to live or on a schedule or API call. Today’s advancement is that we now support automatic cache invalidation – when the data cached has been changed, we automatically flush it from the cache. This innovation means customers can cache dynamic data that’s never been safe to cache before. Never before has anyone offered this type of cache invalidation for SQL Server, MySQL, and Oracle databases.
Imagine the shopping cart data on an eCommerce site. Every time the customer clicks to a new page, the cart has to make a call to the database to populate the shopping cart correctly. That data doesn’t change with every click, but pages pull that data all the time and whenever the customer goes to the cart, the data must be accurate.
Time to live or a schedule would never have worked for expiring shopping cart data, so eCommerce sites haven’t risked caching the data, despite the frequency with which it’s accessed. With ScaleArc’s innovation, any change to the shopping cart automatically invalidates the cache for just the impacted object, delivering the performance benefits of caching this frequently hit data without compromising data integrity.
An eCommerce company extensively tested our new auto cache invalidation technique. The company went from being able to serve 50K queries per second (QPS) to 330K QPS, and response time improved 6x to 12x throughout the testing.
Best of all, unlike other caching technologies, the ScaleArc cache – like all our database load balancing technology – is fully transparent. That is, you don’t have to change anything in the app or the database to use our cache. Another critical tenet of our architecture is to never sacrifice data integrity for the sake of performance – so we don’t cache anything automatically. Our analytics make it easy for you to identify query patterns that are good candidates for caching. And our innovation cache invalidation now lets you extend the data sets you can safely cache, increasing performance for your apps.
comments powered by Disqus