Prioritizing pain points and headaches: or, what's really important in storage.
Key Issues Facing the Storage Industry Today
1) A much easier way to implement and manage SANs (SANs are highly desirable but remain resource consuming and complex to implement and administer).
2) A single storage subsystem that handles files and blocks transparently.
3) Vendors who can tell me what the product means to my business, not just how the product works. This translates into presenting a quantifiable value proposition.
4) How can I determine the value of my data?
5) Dramatically improved, bulletproof storage/network security (as intrusion/hacker threats rapidly mount).
6) The end of SPAM.
7) Non-disruptive storage and system changes, maintenance and upgrades.
8) SMS functionality for Open Systems.
9) A single, clear ILM message and an end-to-end ILM solution from one vendor.
10) Pro-active SRM products (versus reactive).
11) Improving data recovery performance.
12) Storage vendors that scale capacity and performance, not just capacity, and can do it non-disruptively.
13) Achieving compliance without a major impact on budget or resources. Is it worth it?
14) Device to device data movement (without all data movement and migration to new devices going through the servers).
15) Virtual tape for non-mainframe systems.
What's Behind the Issues?
There were many additional issues raised, and the overall list is exhaustive; however, the list above represents the most frequently mentioned concerns. Let's look deeper at some of the commentary and see what they are telling us.
1) SANs offer a management approach that should make managing data and the storage infrastructure much easier. SANs provide additional levels of data security and offer higher availability in case of device failures for both disk and tape subsystems. However, implementing a SAN usually involves testing and integrating numerous software products and possibly heterogeneous (different) hardware products form various vendors. Because of this, over 65% of SANs installed today remain homogeneous (the same) operating systems. IP SANs were viewed as less expensive and interesting but the SAN implementation challenges remained the same.
2) The cry for a single storage subsystem to handle file and block accessed data is growing. Having one storage subsystem such as NAS to handle file data, and disk arrays to handle block-accessed data is costly. Each subsystem requires its own management software and possibly different personnel, adding additional expenses. One subsystem that can handle file and block data without performance impact has become highly desirable.
3) Too many vendors do a good job of describing how their product works. Unfortunately, they come up short in providing a tangible value proposition that quantifies what the product actually means to a business. Faster, more scalable, and cheaper aren't meaningful value propositions unless the benefit is quantifiable.
4) Responses indicated a desire to know the value of data in order to implement the appropriate and most cost-effective backup/recovery strategy. For example, mission critical data should be mirrored but mirroring all data doubles disk costs and is not cost-justifiable, particularly in tough economic times. For most data, a point-in-time copy to tape or disk remains acceptable.
5) As the threat of hackers and terrorists increases, the requirement for security is escalating. Today's security systems are too vulnerable to make IT executives feel comfortable. Storage security has become the newest storage management discipline.
6) The end of SPAM appeared as a headache on everyone's list. Industry estimates say that spam accounts for well over 60% of all e-mail traffic. Managing the spam problem takes a resource commitment.
7) Non-disruptive maintenance and system upgrades are highly desirable
as the industry moves to a point where zero downtime is becoming mandatory. Vendors were encouraged to add non-disruptive maintenance to their hardware and software offerings and to their value proposition.
8) The fact that many people described the need for an SMS (Storage Management Subsystem) for open systems comes as no surprise. The goal to have a single interface and syntax to manage storage for Unix, Linux, and Win2K systems seems universal as too many systems translate into too many software products needed to manage the systems.
9) The topic of ILM is presently a popular one but each vendor delivers a piece of the total solution. Like SANs, the customer is faced with integrating the pieces from a variety of sources adding to implementation complexity. For example, some vendors provide some or all of the hardware need to manage data through its lifecycle, others offer policy engines and still others offer the data mover component.
10) SRM products have provided much needed insight into the storage and network infrastructure. The next step beyond posting alerts for SRM products is to take actions automatically based on user-defined policies without requiring human interaction.
11) Backup is important, recovery is everything. Every minute that a critical application is unavailable is costly and, therefore, faster data recovery reduces financial losses. Improving recovery time should be a key piece of a vendor's value proposition.
12) Vendors often state scalability in terms of storage capacity in their value proposition. This becomes a pitfall, however, as adding more capacity to a subsystem without increasing its performance actually creates bottlenecks. Scalability includes both performance and capacity.
13) To say regulatory compliance is a distraction for business and technology executives is an understatement. To further support this issue, according to Information Week (May 17, 2004, p.42), a third of executives interviewed said that compliance has a negative impact on productivity. The majority of those interviewed said that they don't think they get any benefit from meeting government mandates. Another said forcing compliance on everybody is like giving everyone a root canal when only a few have bad teeth.
14) Migrating from older to newer technology is time consuming and takes valuable server cycles to move the data from the older device to the newer device. This severely impacts response times and service levels for key applications. The promise of moving data from old disks to new disks or from disk to tape devices, directly bypassing the server, was one of the original and yet to be fulfilled tenets of SANs. As storage pools become larger, the time it takes for data migration is becoming prohibitive. Device to device data transfer is now becoming highly desirable.
15) Virtual tape for non-mainframe systems may come as a surprise to some, but the use of tape is increasing in non-mainframe systems for backup, compliance, fixed content and archival data. In addition, the capacity of tape cartridges has increased faster than the capacity of disk drives and now tape cartridge capacity exceeds the capacity of a disk drive. This means that backing up a full volume, or more commonly file(s), doesn't necessarily fill the tape cartridge. Virtual tape systems that have become a part of mainframe systems that routinely improve cartridge utilization (by allocating multiple virtual tape volumes on a single cartridge) are becoming necessary for non-mainframe systems.
Many issues exist for IT personnel to resolve. The list above represents some of the common ones. It is important to prioritize these issues based on your budget, people and other resource constraints that exist. Once a prioritized list is in place that matches the available capabilities, you should set plans in place to address them. Over the past few years, many business did very little to address these pain points.
As I've written before, doing nothing is a strategy--just not a very good one.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Tape Automation|
|Publication:||Computer Technology Review|
|Date:||Jun 1, 2004|
|Previous Article:||Simplifying storage: combining the iSCSI standard with SAN functionality.|
|Next Article:||Fluorescent multilayer recording is here: watch out, blu-ray!|