Delivering High-Value Data for Cloud, Mobile, and Analytics
z/OS Data and Redis
Business and personal computing continue to shift inexorably toward mobile, Internet, and cloud-based models and methods – even as proven systems of record, such as IBM z Systems™, provide the processing bedrock that powers business worldwide. To compete in this complex and ever-changing environment, enterprises must provide very high-volume data services in real time to millions of users. For these organizations, it is paramount to meet the increasing demand with the greatest efficiency and at the lowest possible cost.
Consider the data challenges that a company like Twitter addresses every day. Twitter’s 288 million monthly active users send 500 million Tweets per day. This in itself represents a very high data volume, and the ways in which the data is manipulated and managed multiply the volume and add layers of complexity. Tweets themselves are highly granular, yet they are threaded vertically in lengthy conversations and replicated horizontally as new users join and conversations diverge and multiply. All of this data must be stored for extended periods of time and delivered immediately upon request anywhere in the world.
The modern data explosion poses a unique challenge for z Systems customers. IBM z Systems have always been superlative at handling big data and mixed application workloads. However, integrating that data, and those applications, to a growing range of devices and distributed mobile apps has been challenging. Even as IBM develops more flexible methods, such as z/OS Connect, to bridge the gap between the mainframe and mobile-, Web-, and cloud-based assets, the demand for more flexible data access capabilities (at faster rates and at lower cost) only intensifies. Traditional database systems such as DB2® and VSAM™ were not designed with today’s data-sharing models or transaction volumes in mind. And MIPS-based pricing models often mean that the marginal cost of accessing z Systems data is greater than the marginal value. As a result, modern data exchange models are increasingly costly.
HostBridge Redis for z/OS addresses these problems specifically, allowing mainframe customers to share and integrate data more efficiently and at lower cost. And by leveraging Redis, a database of choice for global cloud-based business, HostBridge Redis enables organizations to standardize on the same technology that is revolutionizing Web-, mobile-, and cloud-based data architectures.
Use Case Scenario: Claims Processing Company
A use-case scenario will help illustrate how HostBridge Redis will benefit mainframe customers. A national drug-claims processing company processes high volumes of prescriptions. Written by tens of thousands of physicians, physician assistants, nurse practitioners, clinical pharmacists, and others, the prescriptions are fulfilled by pharmacies, hospitals, clinics, and other providers from coast to coast. The data collected by the claims processing company must be shared not only with the health care providers and dispensaries, but also with insurance-carrier partners, government health care and regulatory agencies at the state and federal levels, and with the patients themselves. The demand for data – in near-real time – is enormous.
As requests for data multiply, the claims processing company sees a concomitant rise in DB2 request volume and CICS transaction volume – and in corresponding processing costs. HostBridge Redis for z/OS provides an elegant means to handle much of this higher processing volume more cost effectively by enabling requestors to access the Redis in-memory database running in z/OS. Using DB2 replication and triggers, along with CICS events, the Redis in-memory cache is populated from system of record (SoR) transactions. External requestors can then access data directly from the z/OS Redis database, or they can take advantage of alternative means to receive and use the data – either through Redis’s publish/subscribe capabilities or master/slave replication to their own Redis data servers.
The claims processing company gains two significant benefits by implementing an architecture that includes HostBridge Redis for z/OS. First, they are able to provide access to SoR data in a far more cost-effective manner. This reinforces the value of z/OS, CICS, DB2, VSAM, and IMS as the foundation of the SoR. Second, since HostBridge Redis is zIIP-enabled, they realize additional savings by running their Redis workloads on the zIIP rather than the GP. They can continue devoting high-value GP MIPS to SoR transactions, while using the zIIP to expose or integrate SoR data with the ever-widening array of Web, mobile, and cloud apps.
Note: We wrote the high-level Twitter scenario above after viewing a presentation entitled “Scaling Redis at Twitter,” https://www.youtube.com/watch?v=rP9EKvWt0zo, accessed 3/4/2015.
Statistics are from https://about.twitter.com/company, accessed 3/4/2015.