When compared to an x86/distributed server environment, the operational costs of the power consumption versus IBM’s Mainframe z Systems are half, while the performance is 30 percent greater (using z Systems as an example.)
So let’s think about that for a minute!
X86/Distributed server power consumption is double that of an IBM Mainframe?
An IBM Mainframe system will achieve a 30% greater performance too?
Where’s the catch? Sure – the very word “Mainframe” when used today still conjures up images of huge rooms, vast banks of machinery and lots of people (oftentimes replete with white coats!) pawing at various gauges and monitors. Such images obviously don’t help the Marketing efforts of the so-called “Big Iron” manufacturers.
That said let’s look at some cold hard facts:
- 90+ percent of the world’s top 200 Banks still use mainframes
- Many similar mission-critical applications (think of those used by major airlines/credit card companies) also rely upon “Big Iron” Mainframes
Maybe it’s the fact that despite their “dated” image, Mainframes have been continuously updated and refined to cope with the seemingly ever increasing amounts of the truly explosive growth in mobile and online transactions.
The latest z Systems come with an impressive list of capabilities. z Systems machines can process some 111,000 million instructions per second. They can easily process 2.5 billion transactions a day, encrypt mobile and online banking transactions on the fly, and allow those transactions to be monitored with real-time analytics, allowing users to spot potential fraud and find opportunities within transactions as they’re happening.
The z System itself also supports something called single instruction multiple data (SIMD), which lets one microinstruction operate at the same time on multiple data items. “When you start doing advanced modelling and analytics, you can get an 80% performance improvement because of SIMD.
By amassing data from disparate sources (and encrypting same in the process) and then populating a database designed to turn data from the Internet of Things (IoT) into useful, meaningful information – in real-time – users of this technology will be well placed to exploit the opportunities that will remain elusive to firms relying on the slower, typical server-farms using x86 technology.
Further, a series of application programming interfaces lets developers create apps for the system quickly — (an important selling point for banks that need to add new features to their apps frequently for competitive and business reasons.)
In order to cope with the vast quantities of unstructured data (similar to Datacentrers) that will present itself once firms begin to increasingly interconnect their standard products and services with more and more devices and/or the IoT, firms will need to ensure that the entire spectrum of their systems (e.g. network, storage, API’s, hardware throughput and resilience etc.,) can be handled in a safe, reliable and secure manner.
So, what’s this got to do with Datacenters I hear you ask?
Datacenters have the same need for lower power consumption and increased technical performance. Let’s say that a Datacenter is similar to a bank, Data is the new currency after all, and you need to secure it, you need to access it regularly, you need to make sure that it never gets lost or stolen and you can cater for different currencies. This is exactly what a mainframe can do for a Datacenter!
The latest generation of mainframes can lower your operational costs, increase performance while having one of the toughest security measures available. They now cater for Linux (most popular OS in use today), and not just bespoke operation systems, and they have an impeccable reputation for never having downtime with resiliency built in as standard for practically every main component installed.
No brainer? I think so and with the overall footprint of having the ability to run over 8,000 instances of VM’s on one box it makes a very compelling opening discussion to have over a first meeting.
All this is well and good but did you know you can get even more out of the device with tailored management software and configuration tools?
As if the standard “raw power” proffered by a z System wasn’t enough – there are ISVs who support the IBM ecosystem and specialise in areas such as mainframe data performance and optimisation – further optimising the throughput of various applications and routines. Companies like DataKinetics – a company with 40 years of experience in driving greater performance and value from existing systems.
DataKinetics serves the world’s largest mainframe users with a variety of solutions that achieve significant savings – in time, effort, and money. Recent examples include one of the largest US Banks that implemented a solution that reduces the time taken to process some particularly complex batch jobs from over ten hours to less than one minute!
Think about that. 10 hours to 1-minute reduction in processing time for tasks. Imagine what they could do for datacenter processing tasks, batch jobs across the enterprise and others whilst still enabling the mainframe to service up over 8,000 VM’s!
Another client, a leading Healthcare Insurance provider, saw much more efficient resource utilisation that not only reduced their monthly operating costs by hundreds of thousands of dollars each year, but their system’s optimisation also created significantly more available processing space which allowed a planned system upgrade to be deferred for at least 2 years!
These new generation mainframes also boast more advanced capabilities in application availability, resiliency and disaster recovery than even the most bleeding-edge x86 virtualization software. Further, from a security standpoint, viruses and worms that work through stack overflow conditions in distributed systems simply don’t work on the mainframe.
All of this is achieved with no downtime – not even when upgrading software or trying out newly developed software prior to release. Mainframes are still orders of magnitude more powerful than even the largest virtualised distributed systems clusters. For example, IBM’s most recent z System microprocessors run at 5.0 GHz, while today’s commodity processors typically run between 2 and 3.4 GHz. A single z/VM in version 6.1 can accommodate more than 60 virtual machines per CPU; a fully loaded z Enterprise cabinet can hold four modules, each containing six quad-core processors and up to 786 GB of memory with four levels of cache; and the full system can address more than 10 TB of memory and support thousands of concurrent workloads, in many cases running at (or close to) 100 percent utilization.
Each mainframe operating system (OS) can theoretically host somewhere between 10,000 and 30,000 users, depending on the mix of workloads. The mainframe is also still regarded as the gold standard when it comes to batch jobs — up to 100,000 of which can be run in a day within the same OS.
Without being too advertorial for mainframes it does seem to me that mainframes get a bad rap due the longevity of the word, the misuse of the word in certain films and of course the assumption they cost a lot! Well, they may cost more when compared like for like but they simply are NOT like for like – that would be like comparing apples and oranges. And let’s face it they are in a different league, a league that demands a performance increase while lowering costs – Isn’t this what every company is trying to achieve?
So, as competition among Data Centers/Cloud Providers increases – maybe it’s time for them to take a seriously close look at just how much money they could save by using mainframes instead of x86 server farms.
Less maintenance, less power consumption, the ability to place huge computer resources on the same machine as the data that needs to be analyzed/used to verify authenticity etc., ISVs like DataKinetics focused on driving even greater throughput, and all encrypted on the fly might just enable some of us at least to see mainframes in a new light.
Originally published on Compare the Cloud.
During this time has worked first-hand on major Industry Initiatives both in the U.K. and in the USA – such as TALISMAN, TAURUS, CREST, (the Bank of England’s) CGO, Counterparty/Client/Settlement Risk Reporting, CHAPS, Model A and B type Clearing, Intra-Day Payment Netting, Capital Gains Tax Reporting, Regulatory Reporting, Trading Interfaces (from DOT through to FIX API’s and beyond), Multi-Instrument and Multi-Currency systems, Direct Market Access and Custodian Services.
In short, I have been pretty much continuously involved with various types of FinTech for the longest time.