Abstract
In most antivirus products, the local cache mechanism is implemented for the purpose of reducing the time of the file scan. However, the database designed to store cache is limited, a lot of times, that most of the cache data have been purged during the file scanning since it cannot store too many CRC files in the local cache. Traditional methods of building and forming cache cannot effectively solve the problem of low cache utilization. Whereas a client with a traditional cache mechanism needs to send out multiple queries in order to accumulate a usable cache one-by-one, which causes the waste of database and time. In this paper, we first gather the current feedback data and analyze the data, and then, we explore an effective Two-Step clustering scheme to implement for the optimization of the prefetch cache strategy. Our scheme can avoid the cache being cleared prematurely and increase the probability of cache hits. Extensive experimental results demonstrate that our proposed scheme could save 11.76% time during file scanning.
Highlights
INTRODUCTIONCurrent file reputation services (or, file reputation technology) often use a cache mechanism [1]–[3] to speed up response times
Current file reputation services often use a cache mechanism [1]–[3] to speed up response times
Because the client cache’s generated based on historical queries, a query needs to be sent out first so that the returned item could be accumulated oneby-one and stored in the cache. This process of building up the cache one-by-one is slow, it results in many queries that could not hit the cache, so the queries need be sent over the Internet to the file reputation service though it has a slower response time
Summary
Current file reputation services (or, file reputation technology) often use a cache mechanism [1]–[3] to speed up response times. The cache stored on the local, client-side is based upon queries sent to the file reputation service to get. Because the client cache’s generated based on historical queries, a query needs to be sent out first so that the returned item could be accumulated oneby-one and stored in the cache Needless to say, this process of building up the cache one-by-one is slow, it results in many queries that could not hit the cache, so the queries need be sent over the Internet to the file reputation service though it has a slower response time. The SPSS Two-Step Cluster [27]–[30] Component extends this model-based distance measure to both continuous and categorical variables situations. These companies tailor their marketing and product development strategy to each consumer group to increase sales and build brand loyalty
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
