News
Chris Poelker, FalconStor: IT as a weapon: how to determine the value of virtualisation
July 2008 by Chris Poelker, vice president of enterprise solutions, FalconStor
Although many data centre and storage managers are looking for solutions such as storage virtualisation to simplify operations and streamline the infrastructure, I am finding that most are having a difficult time in figuring out how to qualify the upfront expenditures of buying a virtualisation solution versus the return on investment that is needed to sell the solution up through the management chain. Many upper level managers may be strictly focused on the bottom line, and may perceive solutions that make life easier for IT operations as unimportant. What they are missing, and what operations managers are failing to communicate up the chain, is how virtualisation solutions can have a significant impact on cost savings for the company.
First, let’s look at the business problems companies are faced with:
• time to market for new customer facing applications
• data agility for reporting and data mining applications
• data migrations due to technology refresh or acquisitions
• rapid recovery from corruption, attacks, viruses, or hardware issues
• disaster recovery
• data backup
• data volume growth
• information lifecycle management
• risk mitigation
• regulatory compliance
Larger organisations are throwing millions of dollars at these issues using traditional methodologies to attack the problem - traditional tape backup, array based replication, and tape shipping for DR, and disk cloning for rapid recovery of critical databases. Companies are also spending millions on consultants to figure out how to streamline the business and make the company more agile to beat the competition. But like those cool IBM commercials expose, the deliverables of many of these projects can sometimes be “empty promises.”
Your first step in creating an effective solution to solve these problems should begin with creating strategic policies based on data classification.
Here is a simple 12 step guideline for cost containment within the 21st century data centre:
1. classify the data
2. create policy based on the data classification
3. virtualise and consolidate the server infrastructure (GRID/VMware/Citrix/etc…)
4. virtualise the file systems (GFS/DFS/VFM/VxVM/etc)
5. virtualise storage into pools and class tiers (IPStor, TagmaStor, IBM SVC, etc)
6. place data in correct pool based on policy (FC/SATA/NAS/TAPE/WORM/VTL)
7. use correct protocols per pool (FCP/iSCSI/VI/IP/FC-IP/CIFS/NFS/DAFS)
8. use the correct technology per class (Infiniband/FC-SW/FCAL/IP/VI/SATA)
9. use standards (SMI-S/WEBM/CIM/SOAP/XML)
10. leverage transparency to the applications and users
11. automation
12. documentation
Now, let’s look at how most companies are attacking the problems, and how it can be done better. You need to ask yourself: “Why are we looking at virtualisation in the first place?”
The answer is virtualisation provides the plumbing for a “Services Oriented Architecture” By services, I mean all the things that a COMPLETE virtualisation solution needs to bring to the table to make it useful to the organisation. The whole point of virtualising is to simplify the environment while gaining efficiencies of process and procedures to reduce overall costs - in other words, to put real meaning to the often-used terms of “better ROI”, and “lower TCO”.
Virtualisation at both the compute and storage layers offers many benefits. Virtualising at compute facilitates reducing the physical server footprint, and makes disaster recovery easier for applications. Storage virtualisation creates an abstraction layer between the compute elements and physical storage, which enables storage pooling and classification, host transparent data movement between physical storage devices (fixing tech refresh and making data migration real simple), and since the intelligence can now be located at the fabric layer, it eliminates the need to buy expensive licences for the arrays themselves. You can now buy storage in bulk at lower prices, and do things like data replication between expensive primary storage and lower cost SATA arrays for disk based backup and disaster recovery.
When looking at buying a storage virtualisation solution, you need to be sure the solution is comprehensive and address all the data services needs you are looking to apply. The ability is pool storage and mirror data is only a small part of what the solution should bring to the table. A complete solution should be able to do a lot more in order to become the “data services engine” for the organisation to begin the transition to true “cloud computing”
A sample of capabilities to look for include:
• Heterogeneous storage pooling
• Host transparent data movement for ILM
• Thin provisioning from storage pools
• Capacity on demand
• Both disk and tape virtualisation
• Both file and block support
• Integrated security, encryption, compression, single instancing, data replication, continuous data protection, versioning, and policy creation.
It’s a tall order, but let’s look at what value a comprehensive virtualisation solution should provide to increase total ROI. The lists, under four headings, are extensive:
Operational value
• Heterogeneous storage pooling
• Heterogeneous provisioning
• Heterogeneous continuous data replication
• Continuous data protection
• Transparent LUN virtualisation
• Transparent data migration
• Storage device classification
• Tiered storage
• Bare metal recovery
• SAN-based backup
• Transparent ILM
• User-enabled recovery
• Backup window elimination
• Capacity on demand
Business value
• Improved disaster recovery capabilities
• Non-disruptive scalability
• Risk reduction via integrated encryption
• Long term data retention with less media
• Remote office data protection
• Multi-protocol capabilities for all requirements
• Compliance support
• Increased business flexibility
• Re-use of current assets
• Improved RTO
• Improved RPO
• Improved data services
• Improved time to market
• Improved reliability
Economic value
• Reduced tape media costs
• Reduced tape storage costs
• Reduced tape library hardware costs
• Backup license cost reduction
• Reduced disk storage costs
• Reduced storage requirements
• Integrated de-dupe reduces footprint
• WAN cost reduction
• LAN cost reduction
• Reduces array licensing costs
• Reduced component requirements
• Reduced CAPEX and OPEX
Management value
• Role-based management
• Secure access
• Online recoverability to any point
• Reduced administrative overhead
• Single console management
• Secure remote access
• Reduce vendor lock-in
• Extended life of servers
• Reduced complexity
• Reduce CPU load on servers
• Reduced management costs per GB
• Vendor consolidation
Comprehensive as it is, the list above shows the benefits associated with implementing an intelligent storage networking and virtualisation solution - and an important point to remember is that these value/benefit considerations can be applied to almost any IT project for determining the return on investment and total cost of ownership of the solution.
Although virtualisation can be tricky to get your head round, something in these lists speak to every layer of the management hierarchy. As a result, value/benefit considerations like these can be used as a starting point for operations managers trying to educate upper level management on the virtues of implementing virtualisation to solve real business problems, while saving the company money at the same time.
