IBM Mainframes – Integrating Business Analytics With zBX

IBM System Z Update

• Shipped 800 zBX with 450 blades
• Signed 80 new System z customers
• Encourages new mainframe participants with education and recruitment support
• Has added Windows to the supported platforms on zBX and URM
• Focuses on business analytics workloads with Smart Analytics System 9700 and 7100 alongside its Smart Analytics Cloud
• We expect to see Security Analytics addressed in future through its Q1 Labs acquisition

We recently spent some time in London with IBM’s Ray Berard and his team discussing its mainframe impetus. We’re sure you’d like to read more about this leading server platform.

System Z Market Success

Figure 1 shows our assessment of System z revenue and revenue growth by quarter up to Q3 2011 (based on a rolling 4 quarter analysis) and includes major product introductions along the way. We believe System z has had strong success, although its recent revenues have missed peaks at the beginning of 2007 and the middle of 2008. It suffered something of a ‘perfect storm’ in 2008, with the recession cut-backs coinciding with its customers slow down in spending in anticipation of the new machines in 2009. It has spent massively on development of System z, incorporating its impressive technical advances in the current line-up. System z is the most successful of the dwindling number of alternatives to x86 processors in the server area. It’s also well covered by IBM’s Smarter Computing approach, which we have covered extensively in the past.

In the year to September 2011 IBM claims significant success for its System z and associated offerings. In particular:

• It shipped over 80 BladeCenter Extension (zBX) units
• It shipped over 475 blades into these units
• It signed up over 80 new System z clients (although this includes existing users installing a machine in a new location)

We assess System z Mips growth as 40% and revenue, 37% for the annual period to the end of September 2011.

In addition IBM has enjoyed a stronger adoption of System z by partners. In particular:

• It now has around 7,000 applications supported, including 3,250 for Linux and 4000 for z/OS
• 1,200 new and upgraded applications were added to System z in 2010, when it launched the zEnterprise
• It has added around 120 new ISV partners

IBM has engaged in a number of activities to encourage the development of mainframe skills among the younger generation. In particular:

• It has enrolled 814 schools enrolled in the ‘Mainframe Curriculum’
• 32,941 students from 17 countries participated in its ‘Master the Mainframe’ contests
• It has launched SystemzJobs.com, which connects System z clients, partners and businesses with students and professionals

All of this demonstrates that it is doing more than just holding position with System z in the server area.

Developing zBX And URM Business Analytics Solutions

For those of you not familiar, zBX and URM were part of the original zEnterprise announcement in 2009. IBM’s zBX is a box connected directly to either the z196 or z114. It has a 10GB data connection and 1GB service. Unified Resource Management (URM) capabilities are delivered over the service connection, allowing common systems management between the mainframe and the attached BladeCenter. Currently the attached BladeCenter can have two chassis per rack and up to 14 blades per chassis.
Initially customers were slow to adopt these offerings. We noted at the time that the failure to accommodate Windows would limit adoption. In addition to adding Windows to Linux and AIX as supported operating systems, IBM has also been active in developing specific solutions to increase the value of integrated co-processing. The latest flavour of these has been for Business Analytics solutions. In particular:

• Smart Analytics System 9700 – an integration of hardware, software and services based on the zEnterprise 196 platform for large-scale analytics
• Smart Analytics System 9710 – a smaller packaged based on the zEnterprise 114 platform for lower-cost analytics

IBM has also launched its DB2 Analytics Accelerator, which uses attached Netazza data warehousing appliances rather than zBX; its users report increased speeds of up to 1,000 times and – in one case – a ROI of 4 months. It also offers the Smart Analytics Cloud, which it describes as ‘a private cloud computing solution for business intelligence and analytics’.
As always with these events we came away with many ideas about developments and the messages IBM is promoting, such as application software cost savings for System z over x86 servers in large deployments, the role System z plays in test and development and production environments. Through its deep idiosyncratic technical knowledge goes further than other vendors in promoting its advantages in ‘total cost per workload’

Some Conclusions – Expect More Workload-Specific Mainframes

Despite IBM’s investment and educational push around mainframes, x86 machines continue to predominate. Mainframes remain firmly associated with the Financial sector, where we believe 70% of sales take place. IBM has been expansive internationally recently, selling a handful of machines in Africa, but we doubt whether its System z will be considered by many as a general purpose alternative to x86 – certainly not outside large and relatively big medium-sized organisations. It is right to focus on its advantages in lowering the cost of workloads, where the larger the implementation, the larger the savings will be. It is also right to concentrate on producing tuned solutions in the business analytics area to add to the lead it has here in transaction processing. You should expect to see new areas addressed in future, such as security analytics through its Q1 Labs acquisition. The high-speed connectivity between zBX and the mainframe extends the value of association beyond close systems management to clever provisioning of virtual machines. We’re voted before that vertical integration has replaced the horizontal approaches of the past. In the mainframe area IBM’s maturity makes it a leader in business analytics, which HP and others are trying to emulate.