Network Storage: Optimizing Solutions

Virtualization: Networked storage systems create the challenge of efficiently managing a large number of disks that are grouped together into a single system. The question is how to most efficiently manage storage to maximize utilization and reduce effective cost per GB? Improvements in this area are being enabled by disk-level virtualization that allow large number of disks to be managed as if they were one (very large) disk, enabling storage to be allocated as needed, without having to specify which disk data is actually on. Because provisioning is more efficient, fewer disks are required and less storage needs to be purchased to meet the needs of users.

Pointers for a Sound DR Plan

The Plan

First get all the business heads together and understand the business requirements of different applications
Develop scenarios and then do the business impact analysis (BIA)
Consider DR for operating environments, applications, application data, logs and journal apart from people and other infrastructures
Have a formal DR plan
The plan should also consider internal in addition to external factors
Establish Recovery Point Objectives (RPO) for the business functions
Establish Recovery Time Objectives (RTO) for the business functions
Identify and prioritize the critical business functions and activities

People and Processes

Have a central DR plan with defined roles and responsibilities
Have a shadow resource plan in place for the key executives and managers
Have a DR training plan in place
Have a specified budget for DR
The top management including the CEO should be involved in the DR plan

Technology

The critical data should be backed up and validated
Switch operations to DR sites at regular intervals to ensure that the preparedness level is high

iSCSI: For SMEs, iSCSI offers an excellent value proposition. For larger enterprises that have already implemented FC SANs in their data centers, it can co-exist with the FC SANs and complement them. For example, a bank may choose to run its core banking applications on FC SAN and run all midsized and smaller databases on iSCSI. Almost all organizations will see value in iSCSI SAN as it helps achieve tremendous RoI. Implementation of FC-SAN is costly, confining it to data centers.

iSCSI attempts to solve many of these problems. It leverages on two standard protocols, SCSI and IP, to create another standards-based protocol. The iSCSI protocol has been ratified by the IETF as a standard and, therefore, it attempts to eliminate all compatibility issues.

IPSAN: IP-based storage has attained increased prominence at the enterprise level because of the cost factor. It helps in total network storage consolidation of storage resources at a lower cost and centralizes the storage architecture. It also helps in leveraging existing investments in fiber channel SAN. With network storage business growing at 50%, IDC predicts that IPSAN will garner more than 25% of the global storage market by 2007 end.

In case of NAS, it extends NAS consolidation capabilities to include traditional block applications. The top few upcoming trends in IPSAN would include large chunk of FCSAN migrating to IPSAN, major OSM would come out with IP storage system and low cost FC to IP Bridge would be available.

Issues to be considered by CIOs

CIO's need to have a total understanding of what virtualization can do and what it cannot do

What is the real IT issue they are faced?
Are there traditional solutions that can solve the problem?
If not, then, is virtualization going to solve the problem?
Can they justify the investment?
What about TCO over a period of time?
How much disruption the virtualization solution will cause to their current environment?
(In most cases, there will be an initial one-time disruption during the first-time implementation)
Can they use their existing equipment (hardware, software, processes) on which a heavy investment has been made? The answer to this should be mostly�'yes'.
How flexible or open is the solution?
Will it introduce new or unnecessary complexity of management if implemented? The answer to this should be mostly� 'No'.
And, other issues like architecture, scalability, functionality, management, interoperability, and support should also be considered.

SAN-NAS Consolidation: A strong trend in the storage market is that customers are moving towards an environment where both NAS and SAN co-exist. While one would not want to replace a SAN on which mission critical applications like ERP, CRM run, one would like to add NAS like flexibility to the SAN. NAS gateways are an answer to this need. Having either NAS or SAN or both as an information management infrastructure enables companies to deploy their own business continuity solution with little effort, because networked data can more easily be shared across the enterprise, even between remote sites used to replicate and back up data.

Scale-out-storage: Scaled-out storage systems represent the next generational leap in storage system technology. Scaled-out storage system architecture breaks the system boundary by incorporating multiple storage systems into a common pool of storage that can be managed and accessed as if it were a single system. Current efforts in this area focus on supporting large, compute-intensive environments (eg digital animation, computer aided engineering, analytics, biotech, defense R&D, etc) that would overwhelm the throughput of any single machine. Here again, virtualization technology helps by enabling administrators to manage separate storage systems as if they were one large system.

Buying Tips
Standards: As the storage industry has not yet reached a standard, the best an enterprise can do is to go for products and solutions that support all the prevalent standards and protocols, and is interoperable with products and solutions of the maximum number of storage and networking vendors. In other words, a solution or product must be interoperable with heterogeneous server, operating systems, storage software products, network connectivity elements, and other devices ranging from drivers to switches and tape subsystems.

Scalability: Future preparedness is the key. Today, storage needs are growing at a fast rate owing to a number of factors. This means that not only should one deploy solutions keeping in mind the needs of the next two-three years, but also a solution that is easily scalable. And it must be optimized to meet demands for capacity, new applications, and service levels.

Flexibility: A storage solution must offer ease of integration, installation, configuration and operation. Also, look for the backup and restore speed of the solution. See if the storage software has a centralized cross platform enterprise administration for all platforms from a web-styled interface? Does it have integrated disaster recovery and bare metal restore? A solution must have industry-leading consolidation capabilities, advanced software functionality, availability and data integrity to provide end-to-end information protection.

Gyan Ranjan Swain
gyanas@cybermedia.co.in