LV 1871 And IBM Virtualisation Highlights Q211
- LV 1871 is a medium sized mutual insurer which has consolidated, improved services and reduced costs through advanced virtualisation
- IBM has simplified the way it talks about virtualisation, outlining a series of small to large options with associated benefits
- Claims advantages in the number and cost of virtual machines running on its x86 and RISC servers
- Has learnt from decades of experience in virtualisation from mainframe, through Unix to x86 and Cloud
- Believes IBM will be a fierce competitor with the likes of Dell and HP as users invest in virtualisation and Cloud
IBM Knows Virtualisation Like The Back Of Your Hand
I remember well when the x86 server vendors discovered hypervisors in 2003: VMware became a star and everybody talked about the consolidation and usage advantages of virtualisation. Being an old-stager I could remember the Unix server vendors going through the same revelations 10 years before. In those days adding logical partitions to Unix was a necessary step in making those machines competitive against the mainframe and minicomputers, many of which already supported virtualisation. Server virtualisation was initially necessary to allow the sharing of extremely scarce computing resources: over the years it has come to help improve utilisation and provide the basis of consolidation. In 2011 it has extended beyond servers to storage, clients, networking and is the first stage for many on the journey to Cloud Computing.
Needless to say IBM has been at the forefront of most of these developments (its first mainframe operating system to include virtualisation was introduced in1967), so it was interesting to listen in to a teleconference call in which Inna Kuznetsova (VP of IBM Systems Software) and Alexander Triebs of LV 1871 Munich talked about the latest developments from vendor and user viewpoints. They both made valid arguments for positioning virtualisation as the foundation of Cloud Computing, as well as cost savings through consolidation.
Alexander Triebs Describes IT Improvement At LV 1871 Through Advanced Virtualisation
LV is a Munich based mutual insurer, specialising in life, annuities and disability benefits. It is a medium sized company linked with 8.5k independent insurance brokers, which it needs to keep happy. Alexander is the company’s Infrastructure Project Manager, who is in the process of completing a project to modernise the company’s infrastructure. He sees the requirements of the company’s business processes and IT systems as one and the same – to be fast, reliable, flexible, robust, cost effective and efficient. This is hardly surprising, given that 98% of LV 1871’s business processes are built on IT-processes.
Over the years the company has engaged in many consolidation projects, starting in 2004 to improve its infrastructure. Its data centre at the time had a number of components. In particular:
- An IBM host running a classic inventory management system
- Sun Solaris running SAP Financials, CO and other applications
- Windows Servers in a unit linked to the inventory management system
- Windows servers in the back office (in total it had between 80 and 100 Windows servers)
- EMC Symmetrix and Centera storage systems
- Separate tape backup via Veritas (now part of Symantec) and Tivoli Storage Manager (TSM)
- An external print service provider
A diagram of the connections between servers, storage and tape backup is shown in Figure 2.
The seven-year strategy to change the company’s IT included the migration of applications from the host to IBM Power Systems running AIX, having decided that was to be the strategic platform for the actuarial inventory management system. The company also decided to reduce its dependence on outsourcing and to reduce the IT cost ratio.
Alexander managed to move away from the point-to-point connections in the earlier system by introducing tiered storage under IBM’s SAN Volume Controller (SVC) as a layer under farms of AIX and VMware ESX (running both Windows and Linux guests). A synchronous mirror links the storage between the two data centres, which in fact are housed in the same building. An up-to-date schematic of LV 1871’s IT system is shown in Figure 1. It includes 8 IBM System x and 2 Power Systems. He notes that an IBM Power System can run as many as 20 applications on a single LPAR, whereas VMware is always ‘one application per virtual machine’.
The company has also introduced virtual desktops using terminal services and linking via the VPN and RDP to connect over 100 staff, including its developers.
Alexander reports a number of benefits, which include:
- The external effectiveness has led to an improvement in customer satisfaction from the brokers, based on the improved performance and greater stability of the system
- The internal efficiency has created cost savings, based on increase agility (changes now take hours, instead of weeks) and the tiered storage which reduces costs and frees up the company’s dependence on specific individual storage vendors
He has learnt a number of lessons that he wants to pass on to others from the changes. In particular:
- That by building systems you can improve performance and increase both stability and reliability of IT
- That with virtualisation the costs tend to move from servers to storage, so you need to think about deleting as well as creating virtual machines and about de-duplicating data on backup
- That SANs and LANs will become the essential backbones of virtualisation creating 100% availability, so you need to think about redundancy, split fabrics and Virtual LANs
- For security you need to separate the ‘De-Militarised Zone’ (DMZ) from the LAN by running them on different hardware
It was good to hear from a European medium sized business continuing to use Unix servers, even if it gave up the mainframe in the process. Alexander and LV 1871 are great examples for those who think of virtualisation as a secret science. His next steps will be to move the second data centre 6 to 7 miles away from the first. He also thinks that LV 1871 will become more involved with Cloud Computing, although its has no offering for its brokers yet. When it does it might consider something like the services offered to independent financial advisors set up by FNZ last year and profiled by ITCandor.
IBM Sees Virtualisation Changing The Economics Of IT
Inna talked about the way IBM aims to offer additional ‘business value’ beyond basic virtualisation offerings. In particular:
- She talked of the progression of consolidating resources, managing workloads, automating processes and optimising delivery
- Outlined how virtualisation solutions can improve business agility and staff productivity – the latter by improving the server to system admin ratios
- Pointed to the specific advantages claimed by IBM for its eX5 x86 machines in terms of the number and lower cost of virtual machines supported
- Talked up the PowerVM hypervisor on its Power series Unix machines, especially in comparison to VMware running Linux workloads on x86 machines
- Pointed to reductions in server management, storage administration and labour hours achievable through the use of Systems Director software and extensions beyond that into heterogeneous servers with Tivoli
- Discussed IBM’s reach into highly scalable Cloud solutions with SONAS, Cloudburst and its Systems Delivery Manager
Along with Dell and HP IBM is getting very serious about its virtualisation and Cloud Computing portfolio, which is being backed up by weekly announcements. In addition to LV 1871 its customer references include:
- Euronics (Europe’s biggest buying group for consumer electronics), which uses XIV storage, the SVC and Power systems to manage its storage environment with only 2 staff
- Bryant University in the US, which uses BladeCenter, Systems Director and Tivoli Monitoring for energy management to make significant savings in time and money
- China Telecom, which has improved agility and utilisation, while cutting energy consumption through the use of PowerVM, Systems Director VMControl and SVC.
IBM believes that its customers tend to introduce virtualisation as the first step on the journey towards implementing a private Cloud: it is also an essential element of what it calls ‘Smarter Computing’, which describes as ‘Big Data In Optimized Systems, Managed As A Cloud’.
Some Conclusions – A Full Portfolio From A Seasoned Virtualisation Performer
It’s good to see IBM getting practical in explaining its virtualisation strategy and portfolio, not least because some of its early announcements in the software area were complex and difficult to digest. As users introduce virtualisation they often find that the reduction in server and desktop hardware costs are eaten up by the extra they have to spend on storage, networking and the virtualisation software itself. The added maturity of its Power Systems allow them to offer advantages for those needing significant scale, while SVC can help organise storage as a pool of resources, rather than as a series of point products. Inna’s presentation addressed the infrastructure issues rather than specific workloads, which have been fully discussed by other IBM executives publicly in recent months. IBM is in the process of beefing up its virtualisation story, well timed in a year in which competition is increasing and the virtualisation sub-market is undergoing significant growth.
It is also a testament to IBM’s customer focus that it managed to get such a great reference from LV 1871: Alexander spent his time talking about the discovery, implementation and advantages of virtualisation, rather than praising IBM, which is much to his credit. We will continue to watch both with interest.
ITCandor Acronym Buster
CO – Controlling (part of SAP ERP programmes)
DMZ – De-Militarised Zone
LAN – Local Area Network
RISC – Reduced Instruction Set Computing (chip architectures used in IBM Power and Oracle/Sun Solaris Unix servers)
SAN – Storage Area Network
SVC – Storage Virtual Controller
TSM – Tivoli Storage Manager
x86 – chip architectures using the x86 instruction set produced by Intel and AMD