Category Archives: Storage

How much data is a Zetta Byte?


There has been a lot of talk over the past couple of days on the new NSA datacenter. Disregarding what type of data they could be storing, this is an AMAZING engineering feat when just looking at the size and scope of the data mining they will be accomplishing.

It is estimated that the new datacenter will hold around 1 zetta-byte of information! If these numbers being reported by NPR are correct that amount of data is equivalent to:

• 1000 Exabytes or 100,000 Petabytes (largest data centers today)
• 250 Billion DVDs
• 100 years worth of global communications
• Or 75% of the TOTAL data stored globally today

That is a huge amount of data that needs to be sorted, analyzed and managed not to mention all of the processing power that would go into this type of data mining. As more revelations come out about the monstrosity that is this data center, the data geek in me shivers in amazement.

Leave a comment

Posted by on July 17, 2013 in Storage


Why Small Businesses are choosing to go with a SAN!

In today’s SMB market there is a major debate on when and why to use a SAN when technologies like server level clustering can be taken advantage of.  Traditionally, a SAN meant specialized knowledge; intimidating many small businesses with the prospect of additional hardware to maintain. When talking with my customers, there are three main reasons to choose a SAN over server level or DAS storage: Scalability, Flexibility and Availability.  Surprisingly, you can get all three at the same price as traditional DAS.


Most customers make purchases on not only the needs of the business today, but those they can foresee in the viable future and for most this means capacity.  When building out storage in a traditional direct attached model, there becomes a physical limitation to how far a server can scale.  Once these physical limitations are met, customers are forced to purchase and ADDITIONAL server.  This can be very wasteful, as they now have to invest in processing power and additional over head.  In cases where you know you will grow out of your current server, a SAN provides a much better scale out infrastructure.


Just as it is important to plan for the future you do know about, it is also important to plan for the future that you don’t.  Technology and businesses are ever changing, and it is impossible to know everything that is going to be demanded of your IT department.  IT directors are always challenged with meeting these new demands of the businesses with the equipment they have on hand.  SANs are flexible to meet these challenges without having to rip and replace legacy equipment.  Whether you’re looking to improve performance, update an aging fabric or take on new business processes a SAN can change to meet these new demands.


With the introduction of virtualization software, businesses have started to open themselves up to higher levels of risk.  No longer will an outage just effect a single application, but entire businesses can be shut down by a single controller outage.  Additionally, planned outages around maintenance have become more difficult as infrastructures have collapsed.  The solution has been to create a highly available architecture where there are no single points of failure.  While cluster can allow you to have the same availability, by utilizing a SAN you can now run your servers in an active/active architecture and use all of the equipment purchased.

When it comes to creating a datacenter for the truly SMB, it is important to understand where you are going, how flexible you need your architecture to be and what are the costs to keeping that business up.

Leave a comment

Posted by on July 17, 2013 in Storage


Backups – they’re not just for mission critical


How does your business define what data is backed up?  If you’re not backing up everything, you may be costing the company more than you think.

In the many years I’ve been in IT I have seen almost every scenario there is when it comes to backing up your data.  What continues to amaze me is the fact that with how accessible and cheap storage has become there are still companies out there that are only backing up their mission critical data and forget about other departments and data.

What business managers continue to forget are the numerous man hours that have gone into developing their test/dev environments, regression stacks, and other non-critical data.  While true that the business could continue to thrive in the event of a serious disaster, at what costs?  What is the labor cost that would need to occur to re-build?  Is the extra storage more expensive than having your engineers re-invent the wheel? 

I have had this conversation with all to many customers after this realization was made too late.  Hundreds of hours of labor were lost because the environments were not seen as business critical until the after-effects were felt by the business.

This, I have found, is not an issue with small vs. large companies.  One of my job responsibilities when I was an intern (long, long ago!) was to MANUALLY backup the department’s  regression buckets weekly.  Looking back at this awesome responsibility I am amazed that the business: one – didn’t have an automated system and two – entrusted this role to an intern who at the time didn’t realize the importance of backups.  What would have happen if I forgot to hit the “start” button Friday afternoon and the servers needed to be restarted? 

In a recent study by Pepperdine University, the average cost to a business where total data loss occured averaged around $20,577 versus just $557 if the data had been backed up and is easily recoverable and this does not include network downtime associated with having critical departments down which is estimated to be around $50,000 per hour.

In an age of automation and more cost effective storage solutions, I am surprised backing up every piece of your data isn’t a standard practice.


Leave a comment

Posted by on July 17, 2013 in Storage


The Many States of Disaster Recovery

Throughout my tenure with IBM and EMC, I have worked with many customers that are looking for some sort of business continuity plan to help them weather a true disaster.  What I have found, is that this idea of BC or “Disaster Recovery” can mean different things to different companies and the decisions they make are determined mostly by money and need.  What it boils down to is the simple question: “How long can I afford to be down, and how much data can I afford to lose?”

I’d like to take you through the three most popular designs and why customers are choosing these types of business continuity plans. 


First, Disaster Recovery is the ability to return to business in the event of total data loss at a primary location (Site A).  For a lot of companies in the SMB space, this doesn’t mean immediate recovery.  For these customers, they can and do have the ability to lose a day’s worth of data and can be down for a week or longer as they rebuild their IT.  For these customers, simply getting your data offsite through backups will accomplish this task and it is the least expensive option for true DR.


Data Loss: Any new data since last backup until time of disaster

Recovery Time: 1 week or more, depends on how long it takes to purchase and re-build your data center

Cost:  Least expensive


In this second scenario the business is still going to lose any data that was not backed up before the disaster hit but they now have a faster recovery time.  This set up is great for businesses that have the ability to lose data, but not significant downtime on their mission critical applications.   

Data Loss: Any new data since last backup until time of disaster

Recovery Time:  A few hours on your mission critical data as you recover to your offsite servers

Cost:  Moderate.  Have to pay for hosting and the additional servers and networking.



In my experience, the previous two scenarios are good enough for most SMB customers, but there are those companies out there that are either unable to lose any data and need true synchronous replication or are unable to weather any downtime to their business.  For these customers, SAN level replication is a true requirement.


Data Loss: Any data not replicated, can be a few seconds to a few minutes depending on bandwidth and data change rate.

Recovery Time:  Immediate.

Cost: Highest.  Customers in this scenario will have two fully functional IT data centers.  This cost can be minimized by running in an Active/Active configuration.


With the many different options out there for business continuity, it is important for businesses to truly understand the business implications of a true disaster and what their individual company can afford to lose.  Too many times companies are choosing the most expensive option (SAN level replication) because that is what their peers or IT professionals are telling them is best, and not taking a serious look at what their business needs are.  Understanding these needs will help your company in designing a business continuity plan that is right for your needs.

Leave a comment

Posted by on February 15, 2013 in Storage


The emergence of Big Data, and what it means for consumers

Do you feel like you’re being watched?  Well,  you  are.  In today’s digital proliferation, companies are becoming smarter in the way they research and understand their consumers.  Notably, we can see this online, while surfing, shopping, networking, or just emailing.  Have you noticed how the advertisements that pop-up seem to match your interests?  Just how did Amazon know I would buy the next Nora Roberts novel?

The answer lies in IT, and the development of Big Data analytics.  IT was often times considered a money pit, constantly pulling resources from other parts of (your?) business. Data analytics have transformed IT into an integral part of marketing and business development.  Gone are the days of mass emails and blanket marketing.   Companies are now able to collect HUGE amounts of data on their customers’ hobbies, interests, personalities and perhaps most importantly, their buying trends. The data enables highly specialized marketing campaigns, reduced costs and increased  ROI of marketing blitzes.

 So, as a consumer, should we be afraid of “big brother” watching our buying trends and knowing when and how we are going to purchase?  Personally, I like the new model.  I enjoy only receiving commercials on Hulu that I want to watch and getting coupons at the grocery store for products I was planning on buying anyways.  With the proliferation of data analytics and the migration to a digital world, we can only expect our world to become more customized.

Leave a comment

Posted by on February 8, 2013 in Storage, Technology


Buying Used, your guide to a healthy SAN purchase

Since moving into my new role with EMC I have had the opportunity to work with many customers on projects big and small.  One of the larger challenges I have been coming across more frequently are customers purchasing used storage arrays.  While in most cases I am very much against the use of older technologies that someone else is trying to discard, there are instances where buying used can be a great deal.  Unfortunately, in most cases, customers are being taken advantage of and end up spending more money on fixing a SAN not worth keeping than if they had just purchased a new array.

Buying a used SAN is not like purchasing a used car.  Storage arrays, like most technologies, continue to become less expensive with every new generation.  This is because processors, discs, and software continue to become commoditized.  Meaning newer technology can be less expensive and provide you better performance and reliability than the previous generations; making acquiring newer technology easier every year.

Good Reasons for buying used:

  • You are building out a small test/dev department where performance and reliability are not needed.
  • You found a newer array still under maintenance that meets their performance and capacity requirements (go ahead, save some money!)

Bad Reasons for buying used:

  • You believe you are too small to purchase a new storage solution because they are for “enterprise” customers
  • You are trying to reduce spending in your IT department

Buyer’s Guide to a used SAN

  • How many years left does the array have before it is EOSL’d by the manufacturer?  Once the array goes EOSL, the manufacturer will no longer be making replacement parts which means if anything breaks finding replacement parts will become more difficult.
  • Does the array come with all the “pieces” to make it work?  On many SANs, including EMC, we store our OS on the first couple of drives.  You would be surprised how many customers don’t know this and are sold an array without the OS.  This is a huge financial mistake and can cause the array to be useless.
  • Is the array still on maintenance?  Like health insurance, if you want to cover your IT equipment for “what if’s” it typically needs to have been on continuous maintenance.  If the maintenance has expired you will need to pay to have the array examined and re-certified.  This, again, can be expensive.
  • And finally, does the array come with the performance and capacity you are looking for?  If not, you will need to purchase additional drives either from the manufacturer or used third party (used drives can and will fail more often).  In this instance, you can usually purchase a new array for around the same price.
Leave a comment

Posted by on February 5, 2013 in Storage


Private Clouds: Modular, Converged or Reference?

The Introduction of Modular Private Clouds

When the financial bubble burst and many companies found their cash flow diminished, many CIO’s were tasked with the very challenging task from the rest of their management team: “Can you do more, with less”? From this need to continually grow your IT infrastructure with less funds came the major IT transformation that we are seeing today: the move to the cloud.

While the cloud provides many IT companies huge financial advantages, there is still a place in IT where there is a need for internal and local services. So CIO’s started looking into ways to take advantage of this new “cloud” infrastructure with the same benefits of keeping everything in house, enter the Private Cloud.

For many IT professionals the “Private Cloud” can mean many different things, so let me just really quickly go through my thoughts. A private cloud is the virtualization of every piece of your IT architecture: servers, networking and storage, such that you can utilize resources on demand and bill users, or in this case business units, for their usage. The initial implementation of these private clouds were still very modular and still create separation in your “IT as a Service” so that IT admins were still spending a lot of time dynamically allocating these resources for users.

What is Converged Infrastructures?

Converged Infrastructures are the next big move into private clouds and basically provide IT in a box: servers, networking, storage and virtualization. The infrastructure comes pre-assembled and ready to use within 15 minutes of plugging it in taking implementation of mission critical applications from months to hours and companies are now able to reallocated IT resources on projects that help create value and revenue for the business instead of on bringing up their systems.

What you get is an Optimized Infrastructure, Rapid Deployment, Lower Costs, Simplified Management and Singular Support. In all, an infrastructure that no longer needs a large IT staff to manage, so you can reallocate those resources to other parts of the business.

While converged infrastructures seem like the optimal IT solution for many companies, they can be very inflexible. Most companies and IT departments have already made investments in technologies that are not included in the converged infrastructure and either still have this technology on the books or have invested in education for their employees that will no longer be applicable. For these companies, moving to a true converged infrastructure can be overly costly and necessitate additional training.

How then are companies able to take advantage of their current infrastructure and skill sets while still receiving the benefits of a converged infrastructure? Many are turning toward Reference Architectures.

What are Reference Architectures?

Reference Architectures provide the major benefits of a truly converged infrastructure where it is a fully qualified, tested, pre-configured system that you know is going to work right out of the box, but with one major benefit: they are based off of the technologies you use in house. And, in many cases these are stood up by the partner community so that they come pre-assembled and ready to deploy.

So which is right for you?

As with all things, it depends. Whether you are content with your current architecture and like having complete control over the management and configuration of your IT environment or love the idea of having a system that has taken the guess work out of IT or just somewhere in between, there are solutions out there to help you migrate your business to the private cloud and start doing more with less.

Leave a comment

Posted by on February 1, 2013 in Storage