http://www.sybase.com/detail?id=1087602
CoreLogic loans personally started in 1983. We have the largest mortgage database in the United States. We cover 80+ percent of all mortgage data in the United States. On the security side we cover more than 95%. Most of our customers are large financial institutions heavily involved in the mortgage industry. Our users are running ad hawk reports in our system; any field is a fair gain. In the past we have allowed them to run reports across millions of records but over the year but that millions of records have grown to billions of records. Now those same reports that they used to run in the past are running across a lot more data and are taking longer. As we were rolling out new features our system performance became slower and slower, our customers while they were very excited with the new features, they were quite unhappy with the response with the response time of our system. Our SQL environment couldn't keep up with the new features that we were rolling out. We were also adding a lot of data, we were also adding lots of customers so the load on our system became quite high for that SQL environment to keep up with. We had to look for an alternative. We found several database technologies out there but we chose Sybase IQ. Sybase IQ's columnar technology allowed us to add fields without having to re-tune our system. It gave us significant boost in quarry performance, it allowed us to compress our Data, safe storage space, it was also certified with our platform, it was a great fit for us. Sybase's IQ multiplex technology allowed us to grow horizontally. We added more users on our system. We were able to distribute that load across multiple nodes within that multiplex environment. With 15.3 now we can take that one step father. We can take a single query and take that and distribute that across multiple nodes. Let me show you how we are doing this now. This is how we have our current production environment set up. We have a couple of nodes that connects to our san storage area; our users are running quires that get distributed between the nodes within their multiplex environment. When a user runs a query it goes through one of the nodes, retrieves the data from the storage area, and returns the result back to the user. Let's assume for a second that this query takes 2 seconds. Our data has grown over the years and let's assume that this query, this query accesses twice as much data that could potentially take a lot longer time as it has to crunch a lot of data. It's running through the same node on another node with similar horsepower; now let's assume that this query is taking 7 seconds. What we are really excited to see in 15.3 is that now we can take that same query and distribute the query load into multiple nodes for the same query that was taken 7 seconds, now we can take that query and divide the workload in multiple nodes, and now we can still get that query back in 2 seconds. So our data volume growth is not affecting us as much as it was before. With this new feature DQP, now we can meet our customer's expectation, still return the same number of reports in a shorter amount of time and meet our customers SLA. Our internal test of 15.3 shows that on average our query performance has increased by five times. Some our longest-running trend reports have increased by 12 times. During out test we saw parallelization of each quarry across multiple nodes within that multiplex environment. With the new features of 15.3 and as we move more towards a cloud like environment, we feel that Sybase IQ would be an even better fit.
Loading...
9:47
Tips and Tricks Sybase IQ for a Columnar World - Part 01 of 05by mlogica505 views
Oh look, ANOTHER Indian company selling our personal information to the highest bidder just like Intellius.
You people are maggots! Time to start digging up YOUR information and posting it on the internet..it is "Public Information" after all.
Drop dead all of you pukes!
shootstrait357 2 weeks ago
the waste company in the world
sunileshwar85 7 months ago