Blog Host: Brett Cooper READ ENTIRE BIO

A discussion on pricing usually starts off internally with a discussion on budgets and goals ending in research and further discussions. There are many strategies involved in getting the best price for your storage platform.

1. Playing vendors off of one another.
• This is an old tactic that plays well in the space -- when you go to the RFP stage, get the information and work the vendors against each other to get the best price. When doing the RFP spec out the pricing and have the vendor list each item by line item pricing. This way you can see what each line-up costs, as well as the total figure. Many vendors today try to do an inclusive pricing option so it is hard to break down the parts prices. Ask for the pricing to be broken down so you can do the math with your purchasing team.
2. Working with channel, distribution and OEMs partners as well as the storage vendor's direct sales force.
• Channel, distribution and OEM partners get good discounts on the same products that you are buying direct, so check around for pricing and know that many of the products that you are buying are being OEMed under different names -- Think Dell and EMC, IBM and NetApp, HDS with HP and SUN. If you have an established relationship with a channel or distribution partner, it is always wise to check with them, as you may be able to get better pricing as well as already having a contractual purchasing agreement in place.
3. Negotiated pricing -- utilize your purchasing team to negotiate the best price.
• This may require you to plan out your needs over a period of time to get the best pricing based on volume.
• Wait until the end of the quarter or the end of the vendor's fiscal year as discounting usually gets better at these times.
4. Check with your purchasing department for any standard contracts that may be in place.
• If you are a government or educational organization, use your scheduled pricing to get the best pricing and terms.
• Many organizations require that a vendor join the list of standard purchasing contracts for an organization. While this process may take a little longer than issuing a purchase order, it works out any of the kinks of working with the vendor community in the longer term.
5. Use analysts and consultants to compare your pricing.
• Many of these organizations can tell you if your pricing strategy is in line with others in your space. In other words, are you paying too much, or just the right amount to get the solution?
6. Buy in preconfigured packages where possible.
• Specially preconfigured packages are usually at better discounting and pricing levels than buying the parts separately.
• Look for training, services and support to be included in the pricing quote -- again make sure the pricing is broken out.
7. Group your purchases together -- storage hardware, software and services.
• Many storage vendors will give better discounts if the purchase contains products and services that can be discounted as a group. Sometimes vendors can package competitive products into the solution as well. An example of this is IBM's Global Services, who can put their purchasing power based on their size to work for you when you need to buy multiple competitive products, but want a great price.
8. Go to the mat and do the competitive research.
• When you buy a SAN switch or HBA, make sure that you are getting the best price by shopping for the products, even if they are OEMed. You can easily find out what these products cost on the open market by asking your vendor for the vendor part number. Then do your research and have your OEM vendor price match or beat the pricing that you have found.
9. Look at vendor leasing and financing options.
• From time to time vendors, can get advantageous financing terms that may be attractive for your storage solution. Check with the vendor's financing team to understand the various options.
10. Deliver your T's & C's up front -- no surprises.
• Go into the negotiation with the integrity that you want to come out of the negotiation with. Be honest and up front with your vendors, delivering timely information at each stage of the negotiation. If there is to be a delay, let your vendors know. You never know when you are going to have to ask your vendor to go above and beyond and if you always have integrity they will be happy to support you -- you are building relationships for life.

I have truly enjoyed my two weeks with the Expert Answer Center and look forward to my next opportunity to share my thoughts with you.

As always, if you have any questions, concerns, or thoughts on this or any of my blog entries please let me know.

Thank you. Brett P. Cooper
Posted by Brett Cooper Manageability -- processes -- expertise
24 AUG 2006 17:26 EDT (21:26, GMT)
Today, I am going to discuss the criteria of manageability -- processes -- expertise as it is related to choosing the right storage platform for your application.

Manageability -- process -- expertise is related to the internal capabilities of the organization to manage the storage platform and the application within the environment. I wrote an answer to a question in October of 2003 for the Storage Expert section of the SearchStorage Website that asked, "How do I simply my storage management process?"

The answer to the question revolves around three P's:

1. Product
2. People
3. Processes

Many storage vendors provide management tools with their offerings. Some of the tools are homogeneous, managing only the storage vendor's products, while others are heterogeneous, managing the storage vendors' and competitors' products. In either case the management solution that is offered should be evaluated along with the processes involved in installing, configuring, provisioning, ongoing management, alerting and repairing. Many organizations only focus on the physical cost of the storage system, depreciating it based on their financial requirements, but this is only a small aspect of the total cost of ownership (TCO) that an organization has to be aware of. There are personnel costs for managing the storage platform on an ongoing based, and this fully loaded cost is what needs to be measured to understand what the costs of the platform will be. Many organizations lay out a capacity-based pricing model that accounts for the physical cost of the device, management cost, power and cooling costs, as well as any organization specific costs.

While reviewing the manageability of a storage platform, look to an end-to-end solution for management and understand the trade-offs in agentless and agent-full platforms. Look at the physical device itself: How easy is it to manage? Can you understand by looking at the device if an error has occurred and how to resolve it? Are the ports on the device easy to access? Are status lights available and easy to review? Are repairs easy to make with or without tools? These are some of the physical management challenges that you need to look at.

On the management software side there are tradeoffs to be made in terms of the installation and configuration of the software. Some packages require distributed agents, while others are agentless. Agentless platforms may be easier deploy in the short term, but they may not offer as many management capabilities or information about the environment being managed that you may require to do your job. Also, ask your IT team to look at how their enterprise management system and the storage management solution fit into the framework for management within the organization. Or, if your organization is a smaller organization, you may want to look at outsourcing your storage platform and have the outsourcer handle the day-to-day management. Either way, you need to have systems in place to handle the management of alerts that come in from the storage platform; creating help-desk tickets, tracking the tickets to resolution, alerting the team, reporting, auditing, asset tracking, and a host of other capabilities that already exist within most organizations' IT departments. Try not to replicate features between solutions, but look for integrations where lower level platform management solutions feed enterprise management systems. Many storage vendors provide platform adapters or integration modules that connect their management tools to enterprise management systems.

Many storage management solutions handle both the physical and logical configuration of your application environment: showing the path from your application server(s) down through the storage infrastructure to your storage platform. Being able to see both the physical and logical environments in detail provides you with the ability to see configuration challenges and understand consumption by capacity trends.

When the storage management solution is integrated with the application, a layered approach can be taken to manage the overall solution, not gust the storage infrastructure. In many helpdesks, the person at the end of the phone is looking at a variety of screens to understand the health of the application, network, servers and storage. When these platforms are layered together, it is easy to see where the challenges are and how to resolve them. Also, when a user calls into the helpdesk, the person on the end of the phone can understand what is going on, versus being told "I don't know what the challenges are -- call back in ten minutes and we should have more information at that time." Each call to the helpdesk costs real $$'s and should not be taken lightly. Not only is there the costs of the call itself and the helpdesk personnel time, but the users themselves are not able to perform their work, meaning lost productivity. Management solutions can enable the overall organization to operate at peak efficiency levels. Today's auditors require these tools to be tested and proven to maintain the business.

Also, look to the storage vendor's management roadmap and how you can influence it to help you better manage your environment. Look to application integration, partner application or hardware support, and customized views based on management roles within the organization as several up and coming features for management platforms that will aid you in not only doing your job faster, but smarter as well.
Posted by Brett Cooper Physical requirements
23 AUG 2006 16:32 EDT (20:32, GMT)
Today, I thought I would dig into the discussion around the physical requirements of a storage platform. Many companies conduct onsite assessments of the physical environment for a storage platform prior to installing it, including the delivery path from the receiving area to the installation location, ensuring that the product will fit through doorways and onto freight elevators. Power and environmental requirements must be carefully planned out including:

1. Adequate floor space and clearance between other equipment in the environment to allow air flow and access

The height, depth and width of all storage systems are available, but what may or may not be spelled out is how much space in front, behind, above, below, and to the sides of the units are required to maintain proper airflow. Most cabinets today require airflow from front to back and on the top, having done away with side based ventilation due to the lining up of the units in rows. In city locations floor space is at a premium, sometimes costing thousands of dollars per month for a single floor tile in the data center. These rising costs have forced the organizations to consolidate their platforms, or in some cases, move their data centers outside the cities to locations where the costs are lower.

2. Power -- conditioned, protected, and stable power from single or dual power feeds from different sources -- circuits.
• Power connector type required -- NEMA specification

Power is one of the challenges for enterprise data centers and smaller IT organizations alike. Getting good, clean, stable power to drive the machinery in IT is a challenge even in the largest cities around the globe. Installing power conditioning systems, battery and gas fired or diesel generators is a great solution, but many organizations may not have the budgets or infrastructure capabilities to support these requirements.

Providing two different circuits may also create another level of protection against power issues. In the end, there is no simple answer to ensuring availability. One solution is creating remote sites and mirroring the complete infrastructure in case of a power disruption. I would suggest that you work with your local utility company and service providers to plan for the requirements of your data center -- you'll be happy you did up front, because making changes to the power infrastructure during or after construction can be costly, not only in terms of budget, but also in terms of time.

3. Air handling and cooling requirements

With all of the discussion around how to cool today's technological innovations in the data center, you need to understand what kind of heat these devices dissipate and plan your cooling system accordingly. Some of these devices can pump out tens of thousands of BTUs per hour. Think about your home heating system, which is rated in the thousands of BTUs per hour, and you can quickly understand that these devices will require additional cooling units. This discussion doesn't apply only to enterprise units for larger companies, as even smaller systems that are installed in traditional IT closets will require additional cooling capacities to keep up with the heat produced.

4. Floor support structure for the weight of the storage system

I have heard stories of storage systems being rolled into raised floor environments and the floor caving in from the weight of the storage systems. From my limited research I don't believe these to be old wives' tales, rather the real-world happenings when older raised floor systems meet the newer, denser storage subsystems. The weight of some of these devices can easily exceed multiple tons when dual or triple cabinets are used to house the massive storage capacities required by some of today's enterprise applications.

5. Cable management

Whether you use a raised floor or the overhead cable management solutions -- I have heard them called whale bones -- you need to adequately plan out the cable management solution so you can easily manage the cabling and replace them as needed. If you are going to be using optical cables, you may want to go with a third party services organization to help you lay out the environment in detail, and provide the right cable chases so any bends in the environment can be planned out and the cables can all be terminated correctly. There is nothing worse than a long run optical cable that isn't terminated correctly.

One last word, watch out for passive patch panels. In the past, these units have been the location of many of the problems in connectivity. As the industry has moved forward, managed patch panels have come to market that resolve many of the issues of the older, passive patch panels.

6. Security -- physical and logical access

I don't remember where I saw the numbers, but some 65% of all IT-related issues are caused by human error that comes from within the company itself. With this in mind, it is important to ensure that the proper level of physical security, including badge access to physical rooms and key locks on storage systems are in place and are utilized. Also, logical access controls such as user ids and passwords need to be changed from their defaults and locked down. Finally, make sure management ports and SNMP are defined correctly and locked down or put behind firewalls or on dedicated networks.

I was at a large telecommunications company a few years ago and did a discovery of their storage infrastructure using a third party storage management product, and I was easily able to discovery their SAN infrastructure and manage it. I could have changed the zoning setup of their switches or shutdown their switches entirely because they never changed the default user id or password on the switches. This is malicious activity that can easily be prevented. I know that it is hard to understand, but most of the malicious activity in an organization comes from within the organization itself. Check with your storage vendor to see what their security requirements are, and work with your internal security team to create some standards and processes when configuring and adding new storage platforms.

7. Safety requirements -- government standards

Check the requirements for code in your location as well as any governing body standards that your organization is required to have in your solutions. These regulations are spelled out in detail and are relatively straightforward to plan for up front. Also, from time to time, these regulations are updated or a new regulation is put into effect. Keep up on the regulations to avoid any challenges in the future.

Many of the storage vendors offer services to help you plan out the installation of your storage platform of choice, and it is important to take advantage of these services up front. These services work off of a tried and true project plan that has been used during many installations, and can help you identify and solve challenges prior to the installation.

If you are not up for the services route, you can review the product documentation, creating a list of requirements for the installation. Each storage product will include the requirements for installation in terms of power, temperature, weight (for understanding if the floor can support the weight -- more on this later), security, cable management and air flow requirements planning out the installation.

I am sure that there are areas that I missed in the physical planning for a new storage platform. Please let me know your thoughts and if you have any questions or concerns.

Thank you.
Posted by Brett Cooper R&D and timescales
22 AUG 2006 17:27 EDT (21:27, GMT)
Next up on my blog is Company Standing and R&D timescales. When one thinks of company standing many ideas come to mind.

1. Solid company with great financials
2. Customer base that sings the company's praises
3. Support structure that's second-to-none
4. Stable product that can fulfill the needs of its customers
5. Partner ecosystem that can fill any void(s) in the company's solution set
6. Analysts and Press that saying good things about the company and its management team
7. Coverage in the geographies where the company does business
8. Qualified channel and distribution partners that you already do business with

All of the above are important to consider when purchased a storage platform for your application. One needs to understand the vendor from the inside out when making a strategic decision. Not only does one need to be concerned about the company's longevity, but their ability to develop and deliver new and innovative solutions to the challenges of your organization.

The idea of Research & Development, or the big "R" and the big "D," are always discussed as vendors trot out their roadmaps and strategic directions during executive briefings at their briefing centers. These meetings are designed to bring you into the fold of the vendor organization, and answer all of your questions and concerns by meeting face-to-face with the management teams and supporting all-stars within the company.

Take advantage of these opportunities to poke around the vendor and see how they run their environment -- do they eat their own dog food? You may be saying, "What is Brett talking about -- eating dog food?" The idea of eating one's own dog food means that they use their own products in test and production to run their business.

You would think that this would be common sense, but, until several years ago, you would have been surprised to see other vendors' hardware and software inside many of the storage vendors' own data centers and IT organizations. How many storage vendors do you think use disk-to-disk backup internally instead of going to tape? I'll bet that many of the organizations in question still use tape and optical to satisfy their regulatory requirements based on their auditors recommendations. There isn't anything wrong with using tape and optical, but the organization is selling you on the disk-to-disk backup using CAS or another WORM disk solution for regulatory compliance.

Now, you ask, "Are they eating their own dog food?" A company I worked at tested their beta email solution in production and easily found the challenges and bugs within their solution. Also, since every employee's mailbox was stored on the production system, we all were able to see how the solution worked, and were able to improve it prior to bringing it to market. This symbiotic relationship between the IT department and the R&D organization served as a conduit for real-world testing and feature and strategy validation.

Back to the R&D discussion, many organizations no longer have the big "R" in research, as there is no revenue-bearing opportunity for doing research for the sake of doing research. And this concept is driving the development organization, or engineering, to deliver features and enhancements based on revenue-bearing opportunities. The challenge for these organizations is delivering an integrated solution when there are many unknowns in the solution. The engineering team needs to examine and document a plan to satisfy the requirements specified by the product management teams. Through this process, challenges are uncovered that will require prototyping and testing, which may take more time that originally thought.

Development organizations strive to agree to timelines based on the information provided to them by the product management organization, whose responsibility is to analyze market trends and customer requirements, and deliver a thesis on what should be built to generate revenue (as well as the organization required to support the product, partnerships and channels to market). The engineering team reviews the product management thesis and supplies back their comments and timelines. The next step is usually a negotiation process, where the two teams engage each other in discussions on the goals and fine points of the requirements document. In parallel, within the vendors organization, additional discussions are going in the support and services processes, channel relationship management, partner management, marketing organization, finance, and legal -- all focused on understanding what the new or updated solution will require in order to move the organization forward creating revenue bearing opportunities and satisfying the customer need -- closing the opportunities. Once the development team agrees, they assign projects to their team including: architecting the product, setting timetables, taking inventory, budgeting and many other activities.

When evaluating a storage platform for your application it is important to understand the development process from your storage vendor, asking them:

• What is their roadmap to produce new solutions for your specific application?
• What is the release process -- Alpha, Beta, First Customer Ship, General Availability, End-of-Life?
• Many vendors will have different names or acronyms for the above delivery process. Ask for a clear definition of each stage, along with the support available at each stage and the expected quality levels -- you may never run an Alpha or Beta in a production environment, but there are some vendors who run these solutions internally to test out their offering's suitability for your environment
• If you are going to run an Alpha or a Beta and provide feedback, it should be worth something to the vendor. You are testing their solution in your environment, and there is a cost associated with your doing this work and the vendor gets great feedback -- what is that worth to you and the vendor?
• How often does the company make or miss its schedule?
• How often does the company ship major releases, and what do they consider a major release of their platform or supporting products?
• Based on my experience, many of the storage vendors are usually focused on a 12-18 month development and delivery cycle for major revisions. These X.x releases are where architectural changes or major features are added to the platforms.
• Check for disk format or file system format changes, and make sure that they don't adversely affect your environment. These changes may require migration of your data to the new platform to take advantage of new features -- laying out the data in a new file system format or doing a full scale migration of the data to a completely new hardware platform. If this migration is required you need to ask if the storage vendor provides these services for a fee or free to help you make the migration seamless.
• How often does the company ship minor releases?
• Many vendors ship minor releases (x.X) on a quarterly basis depending upon how critical the updates are to the customer base. If you have an issue that is being solved within a maintenance or minor release, stay on top of the support and sales organizations and always read the release notes to see what else may have changed. Some vendors even release x.x.X release for specific issues that have been found and resolved.
• What is the company's end-of-life plan for the software? Is it two years after the next release ships? What about the hardware platform? How long will it be supported for the latest software platform? How about application integration support? Partner support?
• All of the storage vendors have standard policies around these questions. Get the answers up front so you know the information and start plan accordingly.
• Are the releases for the platform timed in a release train so all features are tested and supported on day one? If not, how long does it take to fill in all of the holes in the support matrix? 90 days? 120 days? Longer?
• Are partner applications certified and supported when the first customer ship occurs, or is it at a later date?
• Are any third party software or hardware products required to make the solution work?
• Many companies will require new software to be installed for their later revisions of products, possibly including new agents. You need to check this out up front so you go into the testing with your eyes open, especially if you are going to need to roll out new agents, firmware or device drivers to every server in your environment.

These storage products are very expensive and usually start at levels of capacity and cost that are out of reach of many organizations. Many organizations have a need to start small and grow bigger over time. There are storage products that offer a building block approach -- start at a couple of terabytes and grow larger through the addition of addition controller and storage modules. These building block storage subsystems provide many of the features of their larger brethren, including built-in snapshotting, mirroring and replication between like systems and support for different disk types behind the controller (or head). Many of these modular systems are built using industry standard hardware and a customized OS platform, sometimes built on a Linux or BSD derivative that has been hardened to serve the needs of enterprise customers.

One of the great features of many of the modular systems is their ability to support various disk technologies behind a controller, or head unit. Using Fibre Channel and/or SATA disk drives, customers can specify the performance and availability characteristics of their storage platforms utilizing the same operating system with the same commands and management applications. This modular architecture may go against some of the leading storage providers' philosophies of having customized systems and applications to specific environments, but can provide economies of scale for users who do not have to train on several different architectures. There are arguments for both architectures and I can see uses for each:

1. Customization can mean a greater degree of integration, reliability, and performance for a specific environment
2. while the generalized modular approach can mean support for a broader range of application and environments, but may lack some of the features of the customized systems.

According to the marketing from these larger companies, the storage grids are already being deployed in visualization environments and studios that require massive amounts of performance and storage capacities to bring out the next great movie or animated series. These initial deployments are the start of mainstreaming this technology into the database space where larger and larger databases are requiring more performance and scalability than traditional monolithic solutions can provide. There are new features coming out of this technology, including the ability to split a test database from a live database creating another instance for testing or modeling within the grid environment.

I am excited to see what happens over the coming years in terms of the grid architectures and how they are deployed and how integrations occur for data protection and recovery along with management. With a new paradigm comes new opportunities and new solutions.

Let me know your thoughts on this or any of my blog entries.
Posted by Brett Cooper Supportability
18 AUG 2006 19:13 EDT (23:13, GMT)
Supportability -- SLA -- services are next up for discussion. When many people think of supportability they think of calling into a support center somewhere on the globe and going through the level 1-2-3 shuffle. This is one part of supportability, but it doesn't have to stir up emotions of stress. Instead, one can look at supportability through the eyes of the salesperson for the vendor of your storage solution. Your salesperson is your relationship contact, not only for the sale of the product from the onset, but throughout the post sales process -- the customer-for-life program within most organizations. The customer-for-life program is an iterative process-based program in which the sale is not seen as the first point of contact and closure of the opportunity by the vendor; rather, it is seen as part of the relationship-building stage and their promise to maintain the relationship throughout its lifecycle, keeping a customer for life. I first heard of this process many years ago at IBM and took to heart the idea behind the process -- that each step in the relationship with a customer is critical to maintaining a satisfied customer, and a satisfied customer is more likely to purchase additional products from a vendor that they are satisfied with.

Support is not only made up of call (telephone) capabilities, -- also including on-site support as well as remote support for a distributed environment, training and customized services. Each storage vendor has a menu of support services available to provide the necessary support levels that you require. Also, don't forget about the channel partner that you are buying from, they can offer services for you as well that are backed by the vendor.

Additionally, each vendor records the service level of their products in the field via call-home and onsite audits, and can provide you with detailed "real-world" measurements of their offerings deployed in the field -- all you have to do is ask for the information. Further, many vendors promote uptime numbers in their marketing literature, and should be willing to back these numbers up with service level agreements (SLAs). The key in negotiating these SLAs, is working cooperatively with the vendor to understand your environment and their products, and put in realistic numbers based on your organizational goals. If the product is going to be part of the most critical platform in your organization, and you are going to have a single controller storage subsystem you can not expect 24x7x365 availability. This kind of solution would require a dual controller -- mirrored solution (possibly geographically mirrored) to support a meaningful SLA.

Other things to think about in the support negotiation process:

• Is a spare in the air good enough? All storage and infrastructure vendors have supply depots around the globe that supply spares to their customers -- usually via courier or authorized shipping services. Is this going to be good enough for your organization? You may want to look into have spares kept at your facility for high use parts; for example, disk drives and interfaces (NICs or Fibre Channel ports). You may have to pay for them, or negotiate the specifics of the agreement up front, but the more that you prepare for the support process up front, the better off that you will be.
• Take a look at your environment from a variety of angles -- take a physical walk of your location-through with your vendor. The customer service team from your vendor are experts in realizing the challenges within your environment. It is worthwhile to have your environment reviewed by the expert, and you can take the feedback and make changes accordingly.
• Go beyond level 1-2-3 support and get a SPOC. Not the Star Trek right-hand man to Captain Kirk SPOC, but the Single Point Of Contact. Many storage vendors will provide a liaison to support for an organization that will handle all of your support issues directly, and serve as the first and last stop for all discussions. This person is the relationship person for you inside the company and can rally resources when you need them, and reduce issue resolution times drastically. Usually, there is a cost associated with a SPOC, but if the deal is large enough, some vendors will provide this service as part of the deal.
• It does not hurt to ask. Is your organization in need of a special service or support arrangement? Is there a vendor partner solution involved in the overall platform? If so, you need to ask for integrated support up front. Do not just go on the word of the storage vendor or partner: actually test it out by putting a call into the support organizations from each vendor and see what they do to support you. There should be a process for call hand-off and information sharing in place, as well as documented solution guides for integrating the vendor's products together. Ask for these documents and review them as they should be more than 2-4 pages of marketing text. Also, if you have special needs that the vendor has not thought of, now is the time to engage and ask that they customize a solution to fit your needs.
• Take a test drive of the vendor's support team during the testing period. How many times has your organization run a test of a proposed vendor's solution and had an issue during the testing? I would be willing to wager that there has been a small issue or two during every test and there has been a vendor's pre-sales SE right there to jump on it immediately, but how often is that person standing there after the platform has been put into production? Not often enough, so you need to test drive the vendor's support organization up front. Take a look at their knowledgebase on their website, do some searches for specific issues and see if the answers are easy to understand and put in place within the knowledgebase. Put a call into the support organization and see if your pre-sales SE or salesperson is made aware of the issue and how long it takes for a call back: two hours, four hours, more? In all vendor organizations there are processes and systems in place to handle these interactions, the question is: how well are they followed, and can you get the resolution that you need quickly and easily. If you find the support process arduous up front, then you need to consider it in the overall purchase process.
• Take vendor training classes. At one of the storage vendors that I worked for, there was some super secret information that seemed like common sense to me -- the customers who took training classes were 50% to 60% less likely to open up support tickets and their customer satisfaction levels were consistently in the 80% to 90%. Anyone surprised? The reality is that, if the organization either sends an individual for training, or hosts onsite training, then the team will be more adept at communicating and solving issues when they arise. All of the storage vendors that I know offer training classes and certifications. These training classes are the conduit to the storage vendor, and can increase the level of knowledge within an organization quickly. Look at the storage vendor's Website and see what training classes are available, or ask your salesperson and engineer to help you (or your technical team) pick out the right classes to attend. The vendor's sales team may be able to roll the cost of the training classes into the cost of the deal.

Let me know your thoughts on this or any of my blog entries.

Thank you.
Posted by Brett Cooper Stability and recoverability
17 AUG 2006 17:16 EDT (21:16, GMT)
Stability and recoverability are up next on the list for discussion. Stability and recoverability are related to the "what if" scenarios that occur in the planning for the operation of a storage subsystem. From power outages to failed components, each of these variables needs to be understood and tested before a platform is rolled into production.

I was working with a storage vendor a couple of years ago, and during the test one of the vendor's storage engineers handed the customer's test engineer a pair of grounded and shielded clippers and asked him to cut any wire in the storage system. The customer, rather stunned, said, "Ok!" and proceed to cut one of the cables running between an I/O board and the main controller -- all while the storage platform was under a load. Nothing happened, the system kept working without interruption. Then the vendor's storage engineer said, "Remove a controller card." The customer again, did as directed, and the system kept on working, still under load. Finally, after four or five more tests of the like, pulling hard drives, etc. the customer pulled all of the power to the unit and waited a minute. He then turned the unit back on and checked his server. The timeout had not been reached for the test of 240 seconds (six minutes), and the storage subsystem came back online and the test kept on going. After the testing completed the customer checked the test log and there were no data corruption errors to be found. This is a great example of a stability and recoverability test. If the application was to be tested with disaster recovery in mind, then the testing would need to take this into account. What I am stating is that stability and recoverability are two items that need to be tested prior to rolling a storage platform into production based on real-world tests that could occur.

Don't test for things that are not likely to occur. I remember doing some testing in a lab for a customer where we subjected the storage subsystem to simulated rain to see what would happen if the ceiling failed and rain came in. We tested the unit and it worked flawlessly, noting that the rain contacted the outer shell of the system and never actually permeated through. Is this situation realistic, or not? I guess it all depends on the customer and their environment and expectations. On another occasion, I spoke with a military official who explained that most storage vendors plan for disasters to mean that the data center is still available, when he plans for a burning hole. I was shocked to hear this, but given the actions around the globe I can understand his thoughts. He is dealing with the storage platform being destroyed in its entirety, not to mention the security issues that are involved if the subsystem is captured by the enemy.
Posted by Brett Cooper Storage performance
16 AUG 2006 19:01 EDT (23:01, GMT)
Next up on the criteria discussion -- performance. This is one of my favorite areas of discussion, because performance is an area that everyone has an opinion about. Many people believe that only the highest performing system is right for their application. This may or may not be true, but how may organizations actually measure the performance of their current or proposed application and then provide that information in the RFP? I know many that do, but I also know many that accept vendors' claims on performance without baking off the system itself in a test environment.

Every application has a performance profile that can be mapped to a specific storage system and interconnect (protocol) technology from a variety of vendors. There are tools available to look at the performance profile of your current environment, whether a database, email application or user files. The tools are available from Quest, CA, BMC, Precise (Now Symantec-VERITAS) and others. Most storage vendors have tools that can look at the performance of their storage subsystem and let you know exactly how it is performing over a period of time. You can use these metrics to understand if you need to tune the performance of your current system, or improve your platform performance based on the response times that you are seeing through the application.

Many organizations blame the performance of their application on their storage subsystem when, in reality, it may be an interconnect issue, or a poorly tuned server or application. Surprisingly, many organizations have made updates to their server configurations and seen marked improvements in the performance of their overall solution. Little changes in the configuration, such as HBA Queue Depths in a Fibre Channel environment, can have a big difference on the performance of the overall solution, since more data is flowing between the server and the storage subsystem. I always suggest consulting with your storage platform vendor before making any changes to your solution to ensure supportability and stability -- another criterion that we will discuss over the coming days.

In the end, if the performance of the overall solution is not up to the needs of the application or the users, it could be time to evaluate a new platform for your application, including the storage, server and interconnect. You can use your current platform to understand the requirements for your new platform -- laying out the needs in a logical way and discussing the goals with the owners of the application and the users. If you gather the information and pull together an upgrade plan that clearly lays out the goals for the upgrade, then you should be able to make a business case that gets approved for the new platform.

Let me know your thoughts on this or any of my blog entries.
Posted by Brett Cooper Integrating storage
15 AUG 2006 15:25 EDT (19:25, GMT)
Yesterday's blog entry laid out the criteria for choosing the right storage platform for your application. Today, we are going to dig into one of the criteria: Integration.

You may be asking, "Why is Brett starting with this item, when budgeting is usually where we start?" The integration criteria evolved from working with many organizations over the years as a common thread kept coming up with each opportunity -- there was an integration that was required, either with a current or system or application. Some customers talked about how they had an email or database application that was only supported on a specific storage platform, or of an application that required the ability to handle certain stresses, such as the gravity exerted on a storage subsystem during aerial maneuvers. I have found that, if you start planning out the integration criteria up front, many of the other criteria falls out and it is easier to list them out.

Think of your assessment of the storage subsystem as you would any decision. Sketch out a requirements document and put down the criteria in terms of:

• Must haves
• Nice to haves
• and not required

There is a process in a solution company that follows a similar process -- the Marketing Requirements Document (MRD).

For those who are not from the Product Management organization within a company, a document is created called a Marketing Requirements Document that lays out exactly what is needed by a specific audience from a given solution and a bunch of other information. The MRD is provided to the engineering team that then responds back with the ability of the team to deliver, and what is required to deliver the solution. A negotiation occurs and the teams agree on what the course of action is.

I would suggest thinking of this process when you work with your end customers who will be using the storage, whether a Database or Email Administrator, or the user, whose files are on your shared storage. Each of these business owners requires a certain measure of each of the criteria and needs to understand the trade offs in terms that they understand.

Many organizations have evolved this model to a menu system in which the business owner(s) will select from a cost structure (back to the price model again!) to get the features that their application infrastructure requires. This helps with the negotiation process, and can cut down on the time needed to roll out the platform, as the menu system has standardized the process.

But it does not always cover all of the possible permutations. Changes in the environment, regulations, new technologies, or the way the company does business may facilitate a change in the menu system.

Let me know if you have any questions or concerns.

Thank you.
Posted by Brett Cooper Choosing the right storage platform for your application
14 AUG 2006 18:10 EDT (22:10, GMT)
Today kicks off the start of my two week stint on my Expert Answer Center blog on "Choosing storage solutions for your application" and I am excited to share my thoughts and opinions. The great thing about a blog is that there is no right answer, rather an opportunity to share thoughts and opinions, and ask questions of one another. My job, dear reader, is to open the discussion up and share my insights after years of being in the storage marketplace and help get you thinking about the choices that you are making in your enterprise storage platforms. Whether you are selecting a storage platform to run your enterprise's most important database, or a place to store your user's files, there are some key criteria that you need to look at, including:

• Price/Affordability
• Terms and Conditions
• Performance
• Scalability
• Stability -- Recoverability
• Integration with current solutions -- platforms -- standards
• Physical qualities of the solution
• Supportability -- SLA -- services
• Availability of spares
• Manageability -- processes -- expertise
• Relationship with the company -- strategic vendor
• Company standing -- reputation
• Research & Development timescales

Please notice that I did not use any three letter acronyms (TLAs) that permeate most blogs and technology discussions. The reason is simple: TLAs can have many different meanings to different people and I want to stick with real-world applicable terminology that everyone can agree on and carry forward to discussions with others. Have you ever found yourself searching the internet for a TLA only to be more confused when you found the definition than when you started? I have, and I am sure that I am not alone!

Throughout the next few days, I will define and discuss each of the above qualities, and give some real-world examples of each so you can understand the real-world applicability of each of these criteria and how they interplay with each other. When picking the right storage platform, it isn't as simple as saying, "I'll take the fastest, most scalable solution that costs the least." Sounds like a perfect world, doesn't it? Perfect worlds may not exist, but there are ways to get the solutions that you need today without compromising for tomorrow.

Did I miss any criteria that you use to evaluate your storage platforms? If so, email us and let me know your thoughts. Also, let me know if you have any questions or concerns.

Thank you.
Posted by Brett Cooper