Were you to have been asked that question some years ago, the answer most likely would have bee “I’m not sure”....
Striim – for better integration of HPE NonStop transaction processing with modern streaming analytics platforms.
When it comes to the need to implement analytics for the purpose of detecting changes in behavior and the emergence of trends, it’s important to involve mission critical transaction processing solutions. For the NonStop community, it’s always been all about supporting mission critical solutions and for many decades, NonStop’s real time support of these applications has been the foundation stone for many user-driven business and consumer interactions. Whether it’s that time spent in front of an ATM, sending a text, purchasing a railway ticket or simply paying for groceries, there is almost always a NonStop present in the transaction path. There is a reason after all that NonStop systems are called NonStop!
The compute signatures of systems in support of transaction processing and analytics are quite different and the hardware best optimized for each of them are a direct opposite – transaction processing prefers scale-out options whereas analytics prefers scale-up options. Analytics consumes ever more processing power whereas transaction processing benefits from the presence of multiple processors – a simple case of more cores (analytics) versus more processors (transaction processing). Furthermore, while it is a well-known fact that transaction systems work even better when given more memory, analytics platforms benefit most when given access to enormous amounts of memory – it’s all about in-memory processing working best for analytics. With as much talk that there is today about transforming to hybrid IT, many enterprises need to look no further than marrying transaction processing with analytics to see hybrid IT in action – deployments that combine scale-out with scale-up!
The latest development rollouts of NonStop that support the Intel x86 architecture together with running virtualized on hypervisors atop x86 servers bring with it more opportunity to integrate transaction processing with analytics. And the Intel x86 architecture has the added value of bringing not only standardization but performance gains as well – HPE has made it very clear that the new NonStop X family of servers are capable of very high throughput, enough to support some of the biggest networks on the planet. A single x86 server farm running VMware for instance can now support any mix of NonStop, Linux and Windows. This is a boom for CIOs and data center managers looking to standardize their enterprise processing environments around a single architecture.
But how does it work in reality – how does transaction processing on NonStop and analytics on Linux (or Windows) interoperate? How does the data collected on NonStop influence the analytics performed on Linux (or Windows)? Turns out that the methodology used to replicate databases comes to the party; the same way Disaster / Recovery replication products have done to ensure fast switching to backup sites following an outage at a primary site. It’s called Change Data Capture (CDC) and while there are many ways CDC can be configured, for the NonStop community, it’s all about tapping into the log files again, a familiar practice the community knows well.
Of interest, if you search Wiki for CDC information you will come across the following –
Using transaction logs for change data capture offers a challenge in that the structure, contents and use of a transaction log is specific to a database management system. Unlike data access, no standard exists for transaction logs. Most database management systems do not document the internal format of their transaction logs, although some provide programmatic interfaces to their transaction logs (for example: Oracle, DB2, SQL/MP, SQL/MX and SQL Server 2008).
There’s no magic and really, there’s no “special sauce!” And note too that alongside Oracle and SQL Server, there’s specific reference given to both NS SQL/MX and SQL/MP. In Wiki, no less! None of this has been lost on Striim, with the majority of the Striim team members familiar with this methodology; Striim has made extensive use of the access to NonStop SQL as a source for subsequent processing by Striim for integration with the continuous, real time processing, Striim platform. This was the subject of a recent update to the NonStop community given that focused on NonStop and CDC, noting how real-time CDC –
“Complements your existing data integration solutions and upgrades your data architecture to the speed of your business. CDC makes it efficient and fast to capture and distribute transactions from your databases as soon as they occur.”
But what of analytics? Where does the data from NonStop influence an analytics processor? In case you missed earlier discussions on the topic, there are significant advantages to using CDC methodologies even as analytics can be turned around to process what’s happening on NonStop all in real time –
“Moving change data continuously as new database transactions occur enables you to analyze up-to-date information and respond to time-sensitive issues immediately. By using Striim’s log-based CDC capabilities you can minimize overhead on the source systems (extending hardware lifetime) and ensure timely data processing without facing batch window limitations.”
Having a better appreciation for the importance of CDC and of just how beneficial it is transaction processing solutions running on NonStop where analytics performed on the data in real time is becoming increasingly important – and a key differentiation component to better customer engagement – please contact us to schedule a technical demo. Find out for yourself how Striim’s CDC functionality can boost your data architecture by enabling efficient and real-time data integration. And yes, you can also download a free trial of Striim to see its capabilities.
Katherine Rincon | Senior Vice President, Marketing
d 650.241.0680 x215 | email@example.com
Connect with Striim: Twitter | LinkedIn | Facebook | Events