Nick's Data Storage Blog

Computing and Storage Infrastructure Key to Big Savings

Leaders of information technology are facing extreme pressure to reduce both opex and capex spending. When looking into the playbook on ways to get this accomplished, these leaders are first squeezing suppliers and then going after staff reductions. Previously, the deployment of application software would help the business save money, but that’s not as much the case today. 

So where can most organizations save a significant amount of money and increase their service levels? Computing and storage infrastructure.

To understand why this is the best place to save money, we first must do a review of how we got to where we are with infrastructure. Most organizations deployed applications to increase both the efficiency and effectiveness of their company. The strategic part was the application, not the infrastructure supporting it. In fact we were so worried about the application that we often over-sized infrastructure. Over time, this phenomenon caused the inefficiencies we have with infrastructure.

So, it is easy to understand why server consolidation and virtualization have become so popular.  If organizations stop there, however, they will only be getting a small fraction of the potential savings.  Here is a partial list of infrastructure saving techniques:

· Server virtualization · Storage virtualization · Thin provisioning · De-duplication · Writable snapshots · Multiple protocol connectivity · RAID 6 · WAN accelerators · Tiered storage · Data archival · Virtual tape libraries

Let me be so bold as to paint a vision. What if your infrastructure was dial tone for your applications? And you would deploy an application by setting service levels for security, performance and availability. Sound too far-fetched? This is the definition of private cloud computing.  

Getting started is an exercise in the details. You need to document and understand your requirements starting from the application to the infrastructure. 

Email me at nicksblog@adexisstorage.com

Which of these are applicable to your environment?  Which ones should you do first?  What has the biggest payback? Before answering, you must start with the end in mind—data center optimization. This is more of a journey than an event.  It requires a vision and the architecture to support.  

Take the Disaster Recovery Tour

In this digital age, the speed of doing business increases at an accelerating rate. The safety of your critical business data, and the applications that make use of that data, are paramount. VMware disaster recovery solutions provide rapid, simple, reliable and affordable protection in the event of a disaster…protection that enables rapid recovery from planned and unplanned downtime. Take the disaster recovery tour and see what we mean. 

Have a good day,

Nick Cellentani

Visit Adexis at www.adexisstorage.com

Virtualization Video Series

You may be interested in seeing my complimentary, seven-part video series, "Virtualization, Avoiding the Hype," where I discuss real world tips and techniques about implementing and maintaining a virtualized environment. Virtualization topics I cover include data storage, disaster recovery, server hardware, network, the integrated approach, and services.

Have a good day,

Nick Cellentani

visit Adexis at www.adexisstorage.com

When it Comes to Provisioning, Think Thin to win

Thin provisioning reduces storage consumption by allowing planning for consumption at a higher level other than an application. Without thin provisioning, we are forced to plan disk consumption and anticipated growth at an application level. 

 

For example, your inventory control system is currently using 100GB.  You have seen this application grow at 30 percent per year. So, now when it is time to move this application to a new storage platform you do a little math. (In some circles we call this math planning.)  The math goes like this:

 

Current space consumed + (growth rate* time to review again).  In our case, we have 100GB plus 30 gigs per year and we don’t want to fuse with this for three years.  Please note growth rate is compounded.

 

Timeframe	Spaced needed	Over allocated
Day one	100 GB	120 GB
End of year one	130 GB	90 GB
End of year two	169 GB	51 GB
End of year three	220 GB	0 GB

 

In setting up the new array, you give this application 220GB—and therefore you are wasting 90GB at the end of the first year. I am sure some management person would say, “Hey--just review this growth more often and therefore eliminate the over allocation.”  Good thought, but bad things happen if we do not have enough space allocated to the application...like crashes.

 

The disk manufacturers came out with a technique called “thin provisioning” to help us with this problem. In our example, it works by telling the application it has access to 220GB, but it really consumes only whatever amount is used.

 

So how should we best use this technique to our benefit? First make sure you understand the rules or restrictions the disk manufacturer has on this feature. Next do not think this feature relieves you from the responsibility of keeping track of available space.  In reality, you need to do the same type of planning for space, but now it is at a higher level --a level where there are many applications consuming/competing for the same free space.

 

Managing an available pool of unused disk is at a ratio of many to one. Therefore, the ramification of running out of space in the thin world is that you will crash as many applications that are associated with that pool of storage...ouch! 

 

Why not set what your application sees to the limit of what the operating system can address?  Because this would actually allow the application to grow to its next constraint. Be careful if you deploy this type of strategy. Some operating systems will write out super blocks and thus consume space. 

 

Now that we understand that thin provisioning can increase our efficiency associated with used vs. allocated space, what happens when the file system starts to add “thin thinking” to its bag of tricks?  I’ll be reviewing this over the next several weeks.

 

 

Visit Adexis at www.adexisstorage.com  

Data Storage Survey Reveals Interesting Trends

In a Symantec "Stop Buying Storage" survey conducted by Applied Research in March 2009, some interesting new trends were reported in storage procurement and management that organizations are adopting in the current economic climate. The annual "State of the Data Center" survey uncovered that storage utilization had actually dropped from 60% utilization in 2008 to closer to 50% in 2007 even though storage purchases had increased. Since then, new survey results show that organizations still plan to buy more storage but it also appears that they are doing a better job of utilizing the storage capacity they have.

You can see survey details here:

http://www.symantec.com/about/news/resources/press_kits/detail.jsp?pkid=sdcreport

Have a good day,

Nick Cellentani

visit Adexis at www.adexisstorage.com

Data Domain Changing an I/O Paradigm?

Traditionally when we are designing a disk array for an I/O load, we take the total load and divide it by a rate speed for a disk drive to come up the total number of spindles needed to handle the load.

 

In some cases the I/O load out strips the required capacity causing interesting situation.  We have seen where the required capacity might be 100 TB and the I/O load is so intense it causes the number of spindles to exceed the required capacity yielding 350TB of space. This then drives upcost. 

 

Most array manufacturers have this design issue because they all process data in the array in a similar fashion and they all have access to the same drive technology. So what if you could dedupe the data before it hits a spindle? Would this be a better way to get more I/O with less spindle count? The short answer is “yes,” with some caveats.

 

Data Domain originally entered the storage market as a niche player targeted for backup.  They designed for the deduplication of storage, because of all of the copies we all keep in our backup policies. In fact, the first appliances from Data Domain were optimized for large blobs, like the ones that are output from NetBackup. They accomplished this through an inline, real time deduplication process.

 

This process was and is constrained by the processing power (CPUs) in the appliance.   Fast forward to Intel’s new 5500 chip set and what once was a constraint and a hindrance is about to become a technology paradigm shift. I suspect that when the testing in done and the smoke has cleared, we are going to see that, yes, deduping data before it gets to a spindle is faster than not deduping the data before it gets to the spindle. 

 

So now what?  Data Domain is already a great solution to second tier storage...so in the future they might even be a great place for primary storage. But doesn’t there have to be another “gotcha?”  The only one that I am aware of is that the storage controllers in the Data Domain appliance can only address around 36TB of storage. 

 

If you need more I/O and have a storage requirement over 36TB, then you consider solid state disks drives. 

 

 

Visit Adexis at www.adexisstorage.com

 

Cloud Computing...Food for Thought

I recently attended a partner meeting where the main topic of discussion was cloud computing. I have to admit I am still having trouble getting my head around this concept of converging technologies that deliver cloud computing.

The idea is to provide a platform for application deployment where computing, storage and networking infrastructure are dial tone. That is to say, plug in the application and the rest is taken care of. Normal design of computing and storage infrastructure has us taking into account peak loading.

In the cloud, peak loading would be outsourced to a cloud company. You would only pay for the time you used it. Some estimates have savings at 75 percent over today’s infrastructure. This is significant and has many people considering converting applications away from AIX and Solaris.

I know this last statement does not set well with some of our readers, but consider this, speed is no longer an issue. Intel’s new processor line up, Xeon 5000s are so fast that EMC is has announced a Symmetrix with them as the processor. This was unheard of two years ago.

Does history repeat itself? Think back to when our best technology for data transfer between sites was point to point. Could you have predicted that we would have cloud networks that work better and faster than point to point at a fraction of the price?

Can you picture a world where you input the SLA and the cost structure you desire--and the infrastructure configures and manages the rest. Well...that world is upon us.

We are at the early adopter phase of this technology curve, but I like where it is going. Today, I worry about SAN, tiers of storage, sizing servers, clustering, deduping, backups etc. And if you really look close at why we do this, it is about cost. If tomorrow is 75 percent cheaper and I do not have to worry about these items, maybe we can get back to better serving the business.

Visit Adexis at www.adexisstorage.com

State of Enterprise Storage--a Great Report

I don’t know how many of you have seen this report on the “State of Enterprise Storage” from InformationWeek Analytics in conjunction with Byte and Switch. A total of 328 business technology professionals were interviewed—and 51 percent of them said that security was their number one data storage concern...kind of a surprise.

 

Overall, there is just some great information in this report that I think you will find really interesting and valuable.

 

You can download the report here. 

 

Have a good day!

Nick

Eight Ways Data Storage Efforts can Help you Beat Economy Woes

In these tough economic times when IT budgets are being trimmed to the max, there are still ways to stay ahead by saving more. Some activities MUST continue to keep the business alive...hardware and software maintenance, electricity costs, storage costs, tapes and ongoing consumables.

Here are some ways to take these business critical IT initiatives to a new level of cost reduction:

Server consolidation – you can save on electricity, maintenance costs and decrease downtime.

Storage consolidation – you can reduce your total cost of operations by up to 36 percent.

Data deduplication-- you can reduce the amount of disk space you require by up to 70 percent.

Consolidate maintenance and support contracts – recognize a savings of up to 15 percent.

In-source your disaster recovery facilities – reduce risks and decrease costs by up to 10 percent.

Seek third party support to economize.

Use WAN accelerators – you can reduce your amount of required bandwidth by as much as 50 percent.

Archival technology – can eliminate your need to purchase more storage.

I hope these tips help, and if you have additional ideas – I’d like to hear them, so please get in touch.

Have a good day,

Nick

Visit Adexis at www.adexisstorage.com

Know Your Data Storage Costs

Most of my blog readers know I really believe that understanding your infrastructure requirements stack is paramount in successfully optimizing your environment. One requirement that often gets underplayed is what your infrastructure is costing you today.

Let’s take a look at some major components of cost in our computing and storage infrastructure.

Acquisition costs—initial purchase

• Hardware purchase costs
• Software purchase cost
• Implementation costs
• Initial maintenance cost
• Labor costs to perform the acquisition
• Financing costs
• Floor space cost
• Training cost

Administration cost—ongoing day-to-day use

• Hardware maintenance
• Software maintenance
• Internal labor
• External labor
• Migration costs
• Cooling costs
• Power cost
• Downtime costs
• Rev/version level upgrade cost
• Test environment cost

I’m sure you can think of other costs that need to be included. But the important point here is to document and track these costs. In this world of economic uncertainty, we are mainly focused on cost and payback. By understanding your costs, you enable your organization to easily understand and objectively review technologies for cost optimization. Below are some technologies and techniques you should consider as cost reduction targets:

• Consolidation
• Virtualization
  • Servers and Storage 
• Tiered storage
• Storage architecture
• Archival
• Provisioning
• De-duplication
• Administration automation

These cost reducing technologies or optimization areas directly map to:

• Purchase avoidance
• Licensing reduction
• Faster recovery for both DR and BC
• Outages
• Risk avoidance
• Replication circuit costs
• Reduction of qty of servers
• Reduction in data protection infrastructure
• Meantime between provisioning
• Training

Enabling your organization to really save money all starts with understanding—in detail—where you are spending. After that, the direction and the money will take care of itself.

                                                    

Winner of Data Storage Wish Contest

Congratulations to Eric Wardlow (at left) -- winner of the blog contest for his Data Storage wishes. I selected Eric's wishes because they seem progressive...yet in the realm of possibility. His wishes were...

1) Self tagging meta data. Files that automatically gather meta data about who created it, when it was created, why it was created, what its value to the company is and built in, self-destruct business rules (automatic deletion).

2) Unlimited bandwidth. The ability to move huge chunks of data from the home office to oversees offices near instantaneously over the Internet.

3) Complete data management remote control. I want to be able to do my job from my PDA phone anywhere in the world - 3G speeds are not fast enough... we need speed!

Eric has a bachelor's degree in Geology and has worked in the oil exploration business for about 19 years -- many years working with large Seismic data volumes millions of smaller files. He has managed 100 Terabyte robotic archive systems.

Based on the photo he sent us -- I would venture to say that Eric also is a fun guy!

Thanks so much to everyone who did a submission. It was really hard to select just one. I had a great time reading your ideas and I know others did, too.

Have a good day,

Nick  

Visit Adexis at www.adexisstorage.com

 

Data Storage Records Management -- Eating the Elephant

Historically businesses have always had the responsibility to manage and organize their records.  This first began with the disciplines needed to manage a paper system.  Companies created large file rooms and processes to archive the valuable information.  Over time, the file room went to microfiche and less and less paper was utilized.

Now, for the most part, we use paper only when we have to—and that leaves us with most of our business records contained in our information systems. 

For many reasons, we businesses have neglected our responsibility to properly manage these electronic records, but it’s so important to tackle this vital task. Major factors in properly addressing electronic records are: 

• Compliance requirements

• Legal protection

• Application data size management

 So, when the records management project hits your radar, consider eating this elephant with the end in mind...(no, not the end of the elephant)...but envisioning what the solution will look like in the end.  Components of a records management system.

•  Discovery and auditing tools
•  Rules-based engine, data mover, to drive the proper treatment of electronic records.
•  Multiple data containers that vary in performance, cost and permanency
•  Agreement on …the approach

Real World Experience Implementations we have witnessed over the last several years fall into two broad categories. The first is what we call the classic approach, consisting of a committee of business people paired with business analysts to drive the project. Normally, this group is guided by the time-proven methodologies of discover, analyze design, implement and then support.  

During the discovery phase, the group quickly realizes that the project is huge and more than likely multi-year in scope.  Undaunted by the scope, they set off to decide on each records management policy that needs to be put in place. And this is where the project gets bogged down to the point that measureable progress comes to a stand still.  We have witnessed many companies staying in this phase for YEARS.  

Hold on, there is hope and a better way to eat this elephant. Begin with the end in mind.  Many companies are realizing that putting in place the components of records management systems first affords them the ability to start out with globally consistent polices--and then as they analyze each record management requirement, they have all the tools needed to immediately implement.   

This approach places such companies in a much preferable legal position. When it comes to eating the records management elephant...my best advice is of course one bite at a time, but also begin with the end in mind.

Virtual Desktops -- the hot IT topic in 2009

Hi – I wanted to feature some great information on VDI deployments by James Meyers, a senior principal architect with Adexis. James has an incredible amount of server/storage IT experience and was kind enough to pull together “Virtual Desktops – the hot IT topic in 2009.” 

 

This is detailed, valuable information you won’t want to miss. Check it out and let me know what you think!

 

 

Have a great day,

Nick Cellentani

www.adexisstorage.com

Of Stimulus Plans, Game Freaks and Good Books

We are from the government and we are here to help!” So the joke goes. 

 

In this case they are! The good news is the 2009 Stimulus package extended the accelerated depreciation schedule for capital expenditures in 2009. This provision allows organizations to take up to 50 percent depreciation in the first year, thus lowering taxes.

 

In addition, companies are going to get similar rewards for going green. By reducing their energy consumption and carbon footprint. Details aren’t yet available—but that doesn’t mean you should stop your efforts to going green. 

       

Business Opportunity

On a lighter note, I want bring to your attention a business opportunity.  I am a gamer—died-in-the-wool, first person shooter, Halo, Call of Duty freak.  I got started on the PS3 and the Sony network because of the Blue Ray player. Ok here is the business opportunity.  Somebody should upgrade Sony’s network. COD-4 gets connection errors all the time. It also seems that this occurs only when I am having my best game ever.  Can’t someone sell these guys a better platform? 

 

Great Book

On a more serious note, I just read Ram Charan’s new book. “Leadership in the Era of Economic Uncertainty.” This is a well-written and thought provoking guide to weathering our current economic storm. I would recommend everyone in business read it.  Cash is king and those who can manage it until revenue growth becomes king again will survive.

 

Thanks for reading and stay in touch -- and don't forget to check my Magic Wand posting for the data storage make-three-wishes contest. You could win an iPOD Nano!

 

Visit the Adexis web site at www.adexisstorage.com

 

If you had a (data storage) magic wand...

Dream with me for a moment. If we had an infrastructure magic wand and it could magically grant you three wishes, what would you wish for?

Come on. Stop thinking about the network administrator that looks like Jessica Simpson and can manage 10 petabytes of storage. Focus on the top three pains that you would love to get rid of .

Take away those boundaries, friends. These wishes don’t have to actually be POSSIBLE.

My three wishes would be:

• Infinite amount of storage that is has instantaneous recovery, unlimited I/O and is free.

• A self-managed records management system that automatically follows current rules and regulations.

• Universal peace treaty between all administrators. So, yes the Windows guys, Unix guys, mainframe guys and "network guys" would all get along and respect each other.Ok, the DBAs will never get along.) 

Gee...I think I might be on to something here. I’m declaring a contest!!!!

 

Rules:

• You can’t use any of my wishes. 

• You must submit three wishes by March 27, 2009

• A winner will be picked based on nothing but the fact that I like their wishes best--- but they will receive an iPod Nano, and I will dedicate an entire blog post to them.

Visit the Adexis web site at www.adexisstorage.com

Understanding Computing and Storage Infrastructure Requirements

Understanding requirements is always a good thing. If you know what you need, you have a much better chance of getting it. Understanding computing and storage infrastructure requirements isn’t as daunting task as it first seems. It all starts with the business.

Businesses run application software that sit on a scale of criticality. This scale ranges from “never ever down” to “if the system is up in 10 days that’s ok.”  The applications generate two types of data--structured and unstructured.  From my point of view, that data can exist either in or outside of a database.

This requirements stack places pressure on two major disciplines of the computing and storage infrastructure (I would call this CSI, but everyone would start thinking about the TV show). Those pressured disciplines are records management and availability and protection.

 

Records management:  When documenting in this area you are primarily concerned about access and retention. Access is not just about security concerns such who has access; it should also include a concern about dimension of time. Example: When was the information last accessed?

 

Retention also has multiple dimensions how long you should keep information?  How are you going to make sure you have gotten rid of it? A great organization to help companies understand records management is the American Records Management Association...www.arma.org. The ARMA offers a wealth of free guides and audits.

 

Availability and protection: This is the discipline where RTO/RPO is documented for both business continuity and disaster recovery. My definition of business continuity constitutes a local isolated failure--a server dies, a hard drive fails, a network card fails. Disaster recovery means that you no longer have access to your primary data center.

 

All of the above requirements are needed input for analysis phase in designing your computing and storage infrastructure. It is during the design phase that we add in the other requirements like scalability, interoperability and performance items that need further qualification.

 

Finally comes the fun part-- applying the available technologies or techniques that satisfy and/or exceed the design and requirements stack.Once this is agreed upon, the search can start for which vendor produces the best solution.

 

I find it interesting how many businesses try to short cut the disciplines needed to develop the most cost effective computing and storage infrastructure and then wonder why they are constantly running into problems.

 

For a detailed graphic of that will explain Data Protection Optimization, see http://adexisstorage.com/site/pdf/DPOmethodology.pdf

 

     

Today's Focus: Tangible Results

It’s interesting to witness how people and companies are forced to focus during tough economic times.  I think this is a good thing in many ways. Gone are the emotional projects with no measurable payback, and the focus is now on tangible results. The good news is that IT departments have been using technology to reduce costs and prove ROI for decades—by automating the business through the deployment of application software.

In some instances this type of an approach might still make economic sense, but most organizations would be better served if they focused on optimizing the computing and storage infrastructure supporting these deployed applications. Significant cost savings can be achieved by focusing in this area. When I say “significant,” I mean payback periods in days, weeks and months, not years.  Some examples:

• Archiving data—days to months

• Storage virtualization—weeks to months

• Server virtualization—days

• Deduplication on primary storage—instant

• Consolidating maintenance contracts--weeks to months

• In sourcing disaster recovery—months

• Disk-to-disk backups—weeks

• Backup architecture audit--days

• Outsourcing system administration—days to weeks

What should organizations focus on first? Well, that depends on the organization, the environment and their business requirements. Many different factors play into what will deliver the best value.

The real question companies should be asking is “How do we figure it out?” The answer lies in understanding your current environment and it’s costs...and then comparing that to what solutions are possible and their costs.  In addition, you need to consider what the risks associated with implementing are.

 

Here’s an example of two companies, both unable to meet backup windows, ended up with very different cost savings solutions.

 

Company A is in the publishing business that has been in business for over 100 years. It has a formalized disaster recovery plan that includes a contract for a DR site. The company has a CAGR for disk space of 45 percent, and its DR plan primarily uses offsite tape as the recovery mechanism.

 

Company B is a manufacturer for the government—also using tape as the primary recovery mechanism. It has multiple backup environments and a storage growth of 28 percent a year.

 

Solution for Company A

In analyzing this company’s data it was determined that 80 percent of the data was stagnate. Since stagnate data can be treated as fixed content, the most cost effective solution was to put in place a records management system that archived this data and removed it from the backup requirements. This effectively gave the backup environment increased capacity and allowed the company to meet its backup window.

 

Solution for Company B

In this case, it was noticed in the environment review that Company B had very low tape utilization and backup speeds. It was determined that the environment was not being properly administered and managed, causing poor performance.  The most cost effective thing to do was to update the backup software and hardware drivers. After this was accomplished there was an immediate improvement, and the backup window could be met.

 

So you see, solutions can be very different even for the same problem. But the important thing is they could prove real savings.