• Part 1: Storage and the Evolving Cloud
• Part 2: The Seagate Cloud Strategy
• Part 3: Data Protection, Security, and Cold Storage in the Cloud
• Part 4: The Cloud: Keys to Success

Part 3: Data Protection, Security, and Cold Storage in the Cloud

The Data Protection Model is Different

In traditional IT, hardware RAID, RAID 5, RAID 6 now with the higher capacity drives has been the predominant data protection model.  RAID is also used in these cloud infrastructures, but in large infrastructures, it’s typically not a hardware RAID that’s the predominant data protection model.  It’s replication, and it’s at least 3x.

As we develop storage devices over time, we try to include features to help improve RAID rebuild i.e. hot-swapping drives, hot plugging drives, but quite frankly, not all, but a lot of the largest cloud service providers say we replicate, so we can’t fully utilize such features.  So this is another change in thinking we have gleaned from these these discussions.

Storage Security Requirements

We (Seagate) have this great AES, advanced encryption scheme for data at rest, and they (cloud service providers) say,  “we really can’t use it that way, and, we definitely don’t want a key management system that is complicated, especially in this large scale-out.”

We all know they like things very simple.  However, their challenge is when the drive leaves the data center.  There are two primary use cases, one when the drive is repurposed and goes to another site, and the second is when the drive is retired and destroyed. Typically the process is a third party outside of the data center destroys the drive.  At all costs, the data center will protect the data before that drives leaves the site.  The traditional process has been to wipe the drive.  Now, how long does it take to wipe a 3-terabyte drive?  13 hours.  They wipe it at least 3 times.  39 hours, your looking at days.

They said,  “if you guys could come up with a way to erase the data, and do that in less than 39 hours without a sophisticated key management system, then maybe we can do something.”

So we actually came up with a way to do an instant secure erase in less than a second.  It has a lot of value, especially to these large Internet data centers, and it’s consistent with their current security process. Instead of taking 39 hours to wipe a drive to secure the data, they can cryptographically erase the data in less than a second. So, another example or taking the security technology, our encryption technology, and redeploying it to a feature that fits a need of the cloud space.

Customer Driven Cold Storage Devices

We are hearing from a lot of the large internet data centers that their customers are saying, “Don’t delete my data. Now, I may not access it very frequently, I may not access it at all, but I do not want you to delete it.”

For business purposes, and for regulatory requirements, there’s more and more data being accumulated. Yes even in the cloud there is archive data, a.k.a. cold data.  Service providers are very good at migrating data and they want a way to transparently migrate data from warm systems to a cold system, and then back up again – bi-directional automated migration. Tape doesn’t lend itself to that. You talk to the tape guys, there’s a lot of discussion, a lot of debate, but when you talk to the tier one guys, they say,

“We have a tape infrastructure, but what we want to do is migrate that data bi-directionally, and we want to do it seamlessly. We can migrate data very well, but we need a storage device, a system that has an extremely compelling dollar per gigabyte, and from a system standpoint the other really key feature is power, I cannot take a lot of power.”

Yet another need in the cloud space in terms of providing a system, providing a storage device, that has the attributes to support this type of a storage need.

So what advice does Seagate provide cloud builders?  We’ll cover that in the next post on The Keys to Success in the Cloud.

Stay tuned.

Related Posts:

The evolving cloud and how it impacts storage – Part 1 of 4

The evolving cloud and how it impacts storage – Part 2 of 4

Get the full scoop on the HP ProActive Insight Architecture.

• Part 1: Storage and the Evolving Cloud
• Part 2: The Seagate Cloud Strategy
• Part 3: Data Protection, Security, and Cold Storage in the Cloud
• Part 4: The Cloud: Keys to Success

Part 2: The Seagate Cloud Strategy

From a strategy standpoint, Seagate is going to be the market leader and the technology leader, and in large part, we will be doing that by developing strategic partnerships.  Engaging with key partners, and through that engagement, understanding what their challenges and what their issues are, so we can bake that back into the product development process and optimize drives that better fit these applications.

Engagement to Understand Requirements is Critical

As we have these engagements; there are a lot of topics that we try to understand. We try to understand the differences and challenges and how they impact hard drives. We will be taking a deeper dive into each and talk about everything from the type of data center infrastructure to security to cold storage to architectures. It has been eye opening for us in the way things are done differently in this (cloud) space.

Data Center Infrastructure

Starting with the data center infrastructure.  You can have a 50,000 or 60,000 square meter data center, or a very small, modular, container type of data center, and what we’ve learned across the board is that it really doesn’t matter the type. What really matters is the application and the architecture from a software standpoint that determines what storage device is needed for that application, or a given set of applications in these data centers.

Data Center Environment

In terms of environment, this has been a big trend or topic of discussion. The tier one service providers are building their own data centers, and just about all of them are deploying free air or fresh air cooling economizers for power and cooling efficiency.  Power and cooling is probably the number one operating expense in a data center.  They want to operate their data centers with this free air-cooling more days out of the year, as much as possible.  At one of Facebook’s newest data centers in the northwest part of the United States, they did an environmental study.  They increased the chassis inlet air temperature from 25 degrees C to 30 degrees C, and they raised the relative humidity from 65 percent to 90 percent.  Then, they increased the delta-t temperature from 10 degrees C to 20 degrees C.  What this means to the drive is about a 50 degrees C case temperature. That’s getting up there.  We see this trend continuing to occur, and we believe that drives, as well as all components in the system are going to experience harsher environments.

Pushing the Workload Utilization Envelope

In terms of workload utilization, a lot of infrastructures are virtualized; a multi-tenant infrastructure. As we talk to the cloud architects, a lot of them are in the process of revamping their file systems, their software stacks, and they want to improve the utilization of their key components.  They are basically saying, the workloads one year from now will look nothing like they do today.  And that peaks our interest. Think of it this way, whenever a processor isn’t calculating, isn’t processing, they (the service provider) are not making any money. Whenever a hard drive is not reading and writing, the service provider, again, is not making any money.  So what’s ideal for them, what’s nirvana is for the hard drives to read and write all the time, 24×7, no idle time which impacts us, because we use some of that idle time to do background checks, to do drive scans.  So, in addition to drives being used in harsher environments, they are going to be working even harder.

Pushing the Efficiency Envelope

Remember that story I told you about the delta-t study, the delta-t temperature increase?  The delta-t temperature is the temperature difference between the hot aisle temperature and the cold aisle temperature. That delta-t temperature put the temperature of the hot aisle at 50 degrees C, and the hot aisle is where the maintenance is done.  So, management said, we need to be kinder to our system admins, so they’re going to do maintenance in the cold aisle.  That necessitated the need to change, a design change, that had to move all the cabling to go from one end to the other.  So guess what folks were able to respond very quick to that change, be nimble and adapt, and be very responsive to making that change for this tier one service provider?  Those builders, those integrators working side by side, very closely with these tier one service providers. What we have heard over and over again is this local high-touch technical support is critical to success.

So in what ways is Seagate innovating for the cloud?  We’ll cover that in the next post on Data Protection, Security, and Cold Storage in the Cloud.

Stay tuned.

Related Posts:

The evolving cloud and how it impacts storage – Part 1 of 4

Seagate alone has 7 different drive families in the enterprise storage space with 26 capacity points, and 101 model numbers. Everything from SSDs to hard drives, encrypted to non-encrypted, FIPS to non-FIPS…the landscape can be daunting for anyone to navigate.

You might ask:  Why so many options?  Can’t we just have one or two drives that meet the demands of enterprise servers and storage?  This used to be the case when there was simply Seagate Cheetah 10K back in the late 90s early 2000s. Times have changed…enter the realm of:

Tiered Storage

By definition tiered storage “is the assignment of different categories of data to different types of storage media in order to reduce total storage cost. Categories may be based on levels of protection needed, performance requirements, frequency of use, and other considerations,” according to SearchStorage.com.

What “drives” data’s level of protection, performance requirements, frequency of use, etc. is largely the applications creating and delivering that data, as well as the nature of the data itself, and how quickly, and how often it is needed.  The idea being that the less often the data is needed, the more it should reside on lower cost, higher capacity storage.  The most mission critical data being at the highest tier (Tier 0) – commonly called “the SSD tier,” and the least accessed being in the lowest tier (Tier 3), commonly termed “the archive tier.”

To simplify things – consider this visual by Seagate:

This past March, Wes Perdue, Seagate’s  Director of PLM Cloud Strategy at Seagate spoke at World Hosting Days in Europa-Park Rust, Germany.  The topic: The Evolving Cloud and How it Impacts Storage.  We have taken the transcript of Wes’s presentation and created this 4 part series covering:

• Part 1: Storage and the Evolving Cloud
• Part 2: The Seagate Cloud Strategy
• Part 3: Data Protection, Security, and Cold Storage in the Cloud
• Part 4: The Cloud: Keys to Success

Part 1: Storage and the Evolving Cloud

The cloud space is very important to Seagate.  In fact it’s a strategic imperative.  The service providers do things just a little bit different than their traditional IT brethren.  They push the envelope in a lot of different ways, increasing efficiencies, improving costs, and we’re off as a drive manufacturer to fully understand those differences and what opportunities exist to optimize storage devices for this space.

The WW Cloud Market

From a storage growth standpoint, in terms of US dollars in billions, the worldwide cloud services market is approximately 100 billion dollars and in a couple years, it’s approaching 150 billion dollars.  As a hard drive manufacturer, it’s difficult to hard to get our arms around what this means from a capacity or a drive unit standpoint. But, what we know is that these services enable and drive new applications, and those applications need data, and they need storage.

2011 Cloud Markets by Geo

From a worldwide perspective, about 57 percent of the cloud services revenue is in the Americas, 19 percent in Europe, and the rest in Asia Pacific. If you had to look one area with the strongest growth, it’s probably in Asia Pacific.  They have the smallest percentage, but the largest potential.

Connected Devices need Servers

Over the next few years, analysts are projecting 790 million smartphones and 300 million tablets will be sold worldwide. According to Intel, for every 600 smartphones, you need a server, and for every 122 tablets, you need a server.  So you need 1.3 million servers to support those smartphones. You need 2.5 million servers to support those tablets.  That’s 3.8 million servers to support this mobile infrastructure.  And, servers require storage.

Enterprise Capacity Demand

So, how many drives? How much capacity in terms of enterprise drives that went into a cloud infrastructure.  Last year, 2011, 23 percent of enterprise capacity was for a cloud infrastructure. And in a couple years, that’s projected to be 39 percent. Seagate does not contend that all data is going to go into the cloud, because the nature of some of the data, and/or the culture of some companies. In particular, with public clouds, there is just going to be data companies simply don’t entrust to a third party.  They may create a private cloud behind their firewall within their four walls instead.  Still, we don’t know that every piece of data created is going to be in the cloud at some point in time, but need less to say, a good portion of data will be.

The Demand for Storage Devices

What we do know is that this is driving a lot of storage devices.  By the end of the decade, we are looking at a billion hard drives, and over 200 million solid-state drives shipped worldwide.  If we project that more than half of these devices will be in the cloud in some way shape or form, it’s important to understand how cloud service providers do things differently, and design such requirements into our products.

We’ll cover that in future posts in this series.  The next post will cover exactly how Seagate is strategizing around the cloud.

Stay tuned.

Rorke Data uses Seagate Enterprise drives in their RAID, SAN & Archive solutions. Their customers rely on the performance of Pulsar SSDs and Savvio HDDs along with the large capacities they require with the Constellation ES family of capacity-optimized hard drives.
Learn more about Rorke Data – watch this video.

Read more about the successful partnership here.

He was of course referencing the great annual study conducted by IDC and sponsored by EMC, after which he went on to explain how this will relate to SSD use. SSDs will certainly complement system architectures in the overall hierarchy, but Janukowicz further lays the foundation of just how and where that will occur.

And what about the vendor trends around SSD that will help fuel further growth and adoption? Joseph Kovar took on this topic by discussing not only Single-Level Cell (SLC) vs. Multi-Level Cell (MLC) technology in SSDs, but also how the software and controller technology has changed. For example, Kovar noted that many applications now can be made aware of SSDs and therefore take advantage of specific benefits in their use.

At the device level, SSDs have advanced greatly, and work from organizations including the Storage Performance Council and JEDEC have significantly helped with the process of endurance and performance standardization. This is just one area that Rich Vignes of Seagate covered. He also discussed the differences between enterprise and client-class SSDs, from which the ChannelCast itself was titled.

Interested in knowing more? Although the live event is now over, the ChannelCast presentation is now available to view here. Simply register and then you’ll be sent to the download page. I must say it was refreshing to listen in and view the presentation from these industry experts. And I also came away from it with a much better understanding of what we can expect within the enterprise in the year ahead and beyond. I hope you do as well!

After viewing the presentation, let us know what questions or thoughts you have. Do your experiences differ from the presenters? You can post here or at our Facebook page.

Moving on to “where and how” to store your data, it’s a good idea to rate the application usage within your organization. Here are some general guidelines that may prove useful.
HOT DATA: Up until last year, hot data was typically stored on Tier 1 hard disk drives – namely 10K and 15K-RPM spinning media. Today, some of the hottest of that hot data (Tier 0) has been moved to an SSD (Solid State Disk) for instantaneous access.
WARM DATA: When its data that needs to be read and/or written to often and is critically important to the organization, this Tier 1 data is often stored on mainstream 10K and 15K-RPM hard drives. Much more affordable than SSDs, these drives offer the highest form of reliable, consistent performance for high demand applications.
COOL DATA: The biggest pool of data is stored in Tier 2 primarily because of its cost-efficiency. Offering high capacity storage at the most economical $/GB, this is where the data that must be readily available on-line yet is not critically vital to the organization today is stored.
COLD DATA: Data at rest or archived data has traditionally been stored on tape. As the cost of hard drives has declined and the blurring between on-line and off-line access requirements has occurred, some data centers are opting for using high capacity, low power hard drives for their cold storage requirements.

And now for the “when”. Obviously, data storage is a constant and essential component for successful businesses today and data growth is expanding exponentially year upon year. In fact, staying ahead of the storage demand can be an exhaustive and painful effort. But as we know, what’s hot today probably won’t be hot tomorrow, so storage tiering allows you to move data easily between these tiers making you as efficient as possible. Auto Storage Tiering (AST) will make your life easier. Storage magazine’s recent “Storage Purchasing Intentions Survey” found 27% of data storage managers currently use automated storage tiering (AST) and another 32% will evaluate the technology.

Read more about the tiered storage from Ashish Nadkarni of Taneja Group’s podcast or Randy Kerns of Search Storage blog or read IDC’s paper on tiering.

Are you tiering your storage today?  If so, how many tiers of storage do you have? And if not, what’s holding you back?

Perhaps the largest global cloud opportunities are going to come from China. According to this recent story at Technorati, there are plans to build a 7,800 square meter complex and dub it “Cloud Valley,” and it also happens to be a spot where more than 1,000 orders have already been made for cloud servers. In the same story, research firm IDC estimated that China can expect 30% growth in its cloud market this year alone.

Seagate's Wes Perdue speaks at the 2011 International Cloud Computing Seminar

Seagate recognized the opportunity in the region, and recently partnered with key organizations including the Chinese Institute of Communications (CIC) and the Ministry of Industry and Information Technology (MIIT), to co-sponsor a forum on Cloud Computing in Beijing last month. The goal was to bring key think tanks, researchers, ministry officials and academic institutions  together to discuss and learn about cloud computing and storage needs moving ahead.

The forum, called the International Cloud Computing Seminar (ICCS), brought presenters from Seagate including Wes Perdue, product line manager for cloud computing strategy, as well as participants from Google, Microsoft, Dell, and many others. During his presentation, Wes discussed cloud use trends as well as offering insight into how cloud systems are best deployed.

The feedback from the ICCS participants and attendees after the event was that it was a success, and it was the collaborative nature of the event and the ideas generated that made it so.

As the cloud infrastructure continues to develop and grow in other nations across the globe, having similar forums for engaging discussion and debate could prove to be valuable. Ultimately it’s all about answering the question of how the deployment of these global cloud data centers can be made as useful and efficient as possible for the world’s data.