Mainframe development is still critical to many large companies, but the typical enterprise developer knows very little about it. So we reached out to Josh Rogers, President of Syncsort. Founded in 1968, Syncsort is one of the oldest software development companies in the world.
For most enterprise developers, a typical application involves a relational database from Microsoft or Oracle, a middle tier written in .NET/Java, and one or more web sites for a UI. What’s a typical application look like for a mainframe developer?
You describe a traditional 3 tier application. For organizations with large mainframes, the mainframe acts as the third tier, also called the host. The user interacts with a web front end, requests are processed in a processing tier, generally in an application server and finally a transaction is recorded in a relational database. For large enterprise systems that database is often on the mainframe in the form of DB2. There are also many applications where the transaction processing happens on the mainframe as well, generally by a transaction monitor such as CICS.
Most of the transactions consumers complete with banks, telecommunications, insurance and healthcare companies involve a mainframe. This is true whether the customer is taking money out of an ATM, checking a balance on the web, or buying a new phone in a retail store. Mainframes are so prevalent both because they house important business and process logic that has been developed over the last forty years, but also because they are uniquely able to scale to massive transaction volumes, provide unparalleled availability and offer the most proven approach to security.
There are many legacy applications still in use that do not leverage multiple tiers, but are housed entirely on a mainframe. If you have ever checked in for a flight at the ticket counter and seen all of the typing going on, that ticket agent is likely interacting directly with a mainframe application, generally built in languages such as Assembler or COBOL. It is also important to note that the mainframe has a wide range of tooling or JOBS available that can run popular application servers, such as C and Java, and can act as the enterprise’s ‘data of record’ for reporting (financial, operational, etc.), updating various databases on a nightly or periodic basis.
How many mainframes are still in use? And how do you see that number changing in the future?
It is difficult to nail down how many mainframes are actually in use today, but there is no doubting that the mainframe is still critical to many organizations. According to SHARE Inc., an organization focused on educating, networking, and influencing IBM mainframe users, their mainframe community is represented by more than 20,000 individuals from nearly 2,000 companies. Those companies consist of state and federal government agencies, universities, retail, energy, manufacturing, banks, and insurance companies, including:
- 96 of the world’s top 100 banks, 23 of the 25 top US retailers, and 9 out of 10 of the world’s largest insurance companies
- Seventy-one percent of global Fortune 500 companies
- Nine out of the top 10 global life and health insurance providers process their high-volume transactions on a System z mainframe
According to this community, mainframes process roughly 30 billion business transactions per day, including most major credit card transactions and stock trades, money transfers, manufacturing processes, and ERP systems.
Syncsort has a unique view into the mainframe market given our customer base. We price our some of our software based on MSU’s – a measure of processing power deployed – and therefore have reasonable into market trends. We see MSU’s growing about 5% a year and generally see higher growth in larger mainframe companies, which tend to be the largest companies in the world and smaller growth in smaller environments. The value proposition of our solutions is to help customers control their mainframe costs by making their processing more efficient. Our mainframe business has been growing significantly faster than MSU growth demonstrating that customers see their mainframe processing growing and they are looking for ways to control those costs.
Syncsort was originally a mainframe software company, but now you are branching out into many other sectors. Would you like to talk about some of your product lines? (Anything you want, but I think cloud and Hadoop would be particularly interesting to our readers)
Syncsort is an expert at efficient movement and manipulation of data at large scale. We started on the mainframe and moved into distributed platforms when those emerged in the early 90’s and have continued to apply our expertise to data processing environments as new ones emerge. Today, we have taken our technology and integrated it with the latest new platforms gaining tremendous momentum, like Hadoop and Cloud. For Hadoop we offer a high performance ETL solution called DMXh, which allows customers to leverage their existing skills to develop and run large scale ETL routines in Hadoop. Customers want to take advantage of Hadoop for large scale processing but are slowed down by a lack of skills to build the jobs in this new environment. We eliminate the skill challenge and improve the performance of your overall system. We also offer this solution for cloud environments. We can be found in the Amazon Web Services Marketplace where our product can be deployed in just a few clicks on either an EC2 instance or in EMR (Amazon’s Hadoop offering).
Perhaps what is most interesting is the collision between these new Big Data platforms and the mainframe. We see our customers looking to leverage Hadoop or Cloud for analytics but struggle with getting their core operational data into those platforms. We have put a lot of effort into providing solutions to make that easier. DMXh, for example, offers market leading support for pulling and preparing mainframe data into Hadoop.
Another example is a product we released in Q4 of 2014 called Ironstream, which is the product for moving mainframe log data into Splunk – a leading analytics player whose specialty is allowing customers to gain insight from machine log data and has been experiencing explosive growth. Linux servers generate terabytes of log data that can provide unique insight into security risks, IT operations and application performance. They have over 10,000 customers and are nearing $600 million in revenue. We are their solution that allow customers to access mainframe log data. A large mainframe can generate 10 TBs of log data a day. It is hard to get to, is produced in cryptic formats and needs to be moved efficiently off platform without increasing mainframe costs or impacting the mainframes performance. Ironstream handles all of those issues for Splunk customers.
As we look at future opportunities, we are focusing on this collision between Big Data platforms – where people want to analyze data – and the mainframe – where the most important data is produced and recorded. Given our expertise and our technology heritage, we believe we are uniquely able to solve this challenge.
If a programmer was interested in learning mainframe developer, where would be a good place to start?
IBM remains the technology and thought leader for zSystem mainframes and they offer a variety of training programs based on needs. A good starting place would be to check out: http://www-03.ibm.com/systems/z/education/. The SHARE organization is also a good resource for mainframe education: http://www.share.org/