HOW STORAGE TECHNOLOGIES ARE EVOLVING TO MANAGE THE DATA DELUGE Publication – Information Week Dataquest Edition – National Date – March, 2013

Sandeep Dutta Country Manger, Storage India South Asia Exponential increase in data within enterprises is creating a surge in the storage requirements. Vendors are pursuing this opportunity by innovating and launching products around emerging storage technologies like flashbased SSDs, automated tiered storage, storage virtualization and cloudbased storage With enterprises witnessing dramatic growth in structured and unstructured data, analyzing these huge data sets for extracting critical insights has emerged as a top business priority. A recent IDC research report says that extracting value from the expanding universe of digital information is becoming a core business mandate, with organizations spending more time and money collecting, storing, and monetizing fast-growing data sets and rich pools of content for the foreseeable future. In order to enable their organizations to unlock the business value of data, CIOs are confronted with the challenge of designing and maintaining a robust storage architecture that could effectively support the spiraling data needs of the company. This not only involves forecasting and mapping the storage requirements in tandem with the enterprise's growth but also involves managing the entire storage architecture cost effectively and optimally. The fact is corroborated by a recent IDC report, which highlights that the continued expansion of business-critical information and rich content within extended enterprises is today changing the storage dynamic in a wide range of industries. Perceiving the storage-linked challenges that information explosion is set to bring in, technology vendors have been working on developing solutions that could make their storage system offerings not just robust but also intelligent and flexible to fit in specific storage needs of different enterprises. Let's look at some of the emerging trends in storage technologies which are becoming increasingly relevant in today's times. FLASH-BASED SOLID STATE DRIVES Traditionally, Hard Disk Drives (HDD) as a storage technology was widely used in the enterprises. But in the past few years, flash-based Solid State Drive (SSD), which was previously adopted extensively in consumer technology products, has started making its inroads into the enterprise IT environment. The reason why SSD technology started finding application within enterprise IT was because it could deliver much higher performance — that is far greater Input and Output (I/O) data rates — than the conventional HDD. This is because SSD is based on the solid state semiconductor technology unlike HDDs that are spinning disks and lead to noticeable delay in data retrieval. "SSDs that have evolved as flash drives have now created a new use- case scenario in the industry where you can bring in high performance and bring down the cost of data center and power and cooling. This is due to the fact that this technology can handle larger part of the workload, without really increasing the cost of data center operations," says Deepak Varma, Regional Head-Presales, EMC, India & SAARC. Entry of flash-based SSDs in the enterprise IT environment has started becoming highly relevant especially for certain dynamic and high-volume workload industry verticals where certain business

applications need to provide quick query responses with an extremely low latency. "Flash technology has fundamentally changed the paradigm of storage systems and is enabling new use cases for essential enterprise applications and solutions by enhancing their performance, efficiency, and design," says Sandeep Dutta, Country Manager, Storage, IBM India/South Asia. Citing an example of a high volume workload industry vertical where SSDs are increasingly becoming highly relevant, Amit Malhotra, Director, Storage Sales, JAPAC, Systems Division, Oracle Corporation says, “In intense trading environments of capital markets, where every milli-second taken to respond to transactions counts, SSDs are becoming highly relevant as a preferred storage technology to ensure high performance of the trading-linked business applications.” Another example where SSDs are becoming applicable is that of e-commerce websites, where at a given point in time, a large number of users stream in and search for products of their choice. In such a scenario it is essential for the business to ensure that when a user puts a search request, the web page containing the results appears without any noticeable delay. And ensuring such a quick response time means that the data need to be read from the back-end storage at a very high speed, which can be done effectively by using SSDs as a storage medium. “Solid State Technology will continue to be applied broadly to accelerate a wide range of workloads, from virtualized servers and desktops to online transaction processing (OLTP) to file services,” adds Deep Roy, Consulting Systems Engineer, Technology Solutions Organization - APAC, NetApp. Understanding the relevance and application of SSDs within the enterprises and the large enterprise storage market that is waiting to be tapped, more and more technology vendors have started focusing their energies in building innovations around SSDs. INNOVATIONS AROUND SSDS Within the enterprise IT environments, there are always certain business- critical applications that need to have quick response time and are used much frequently than many other non- critical business applications. However, ensuring high performance of the business critical applications over the others has been one of the biggest challenges for the ClOs, as both these apps take the same path of going over the network and access the same back-end storage systems. An interesting application of the flash technology that addresses this issue is by using it as a server flash caching solution, which is an intelligent software technology that makes a copy of the hottest data accessed from the storage system and makes it reside within the flash-based SSD sitting in the server. This means that the cache gets located closer to the application, within the server rather than on the storage arrays. ‘If there is a particular business application on the server, which is most hungry in terms of data access or performance, using the server flash caching solution we bring this data that was residing in the storage system closer to the server. This improves the overall response time of the application to the business,” says Varma of EMC Though, flash technology has its performance benefits, one of the biggest roadblocks for the enterprises to actually adopt this technology as a mainstream storage option has been the fact that SSDs are more expensive than the traditional storage technologies. To address this specific issue Hitachi Data Systems introduced Accelerated Flash Storage last year, which is an innovative flash memory controller technology. “We have built our own multi-core flash controller technology and we have integrated the controller with the flash modules under Hitachi Accelerated Flash storage. This meets the customer demands for higher capacity and performance but at the lower cost per bit by being able to increase performance of multi-level flash to degrees that exceed that of a single flash," says Vivekanand

Venugopal, VP and General Manager, India, Hitachi Data Systems. Another vendor VMware is also innovating in the space and is set to bring an interesting concept of virtual flash to the market."We have virtualized CPI, memory, and hard disks, but there is one thing we have not yet virtualized, which is storage flash memory. At VMware, we are working very closely with some of the storage vendors in the industry to come out with virtualized storage flash memory, which we call as Virtual Flash. This will bring in the benefits of visualization of the flash to the customers," says B S Nagarajan, Director-Systems Engineering, VMware India &SAARC This concept of virtual flash would basically enable the enterprise users to view and manage the flash devices that are plugged into vSphere hosts as a single pool —just in the way other resources like memory or CPU are currently being viewed and managed by the users. AUTOMATED TIERED STORAGE The concept of tiered storage in itself is not new. Within many industry verticals there exists a storage architecture that is differentiated into various tiers, grading from the highest performing storage array — tier zero — meant for active data that is generated through frequently accessed mission-critical applications to the lowermost cost- effective storage tier, which need not necessarily give great performance but proves as high capacity storage system that primarily serves the purpose of storing the enterprise data, which is rarely used. In such tiered storage architecture, a huge challenge that existed within the enterprises was that the IT team had to undergo the cumbersome process of regularly studying the dynamic storage needs of the enterprise, map it against various storage tiers and manually distribute petabytes of data across various storage tiers. “Till 3-4 years back, it was the role of the database administrator or the system administrator to place specific enterprise data in the right storage platforms, which could be different storage media types or different storage arrays to meet the service levels/'highlights Hitachi Data System's Venugopal. This challenge was resolved with the emergence of software that could intelligently understand the pattern of data usage by the business, enabling automated tiered storage. “The software automates tiering of storage by intelligently identifying and prioritizing the hot data, which is used most frequently and by automating the relocation of cold and hot data across the different storage tiers. Automated tiering is highly relevant in the structured data environment, wherein users run queries in a structured way, leading to improvement in efficiency in terms of cooling performance and usage," says Sanchit Vir Gogia, Principal Analyst - Emerging Technologies, IDC India. The primary reason why organizations are betting big on automated tiering is because it is a cost-effective storage option, “High-end storage comes at a high premium, and using top-shelf storage for all the enterprise data assets is no longer a cost- efficient solution. This is increasingly driving organizations to look at intelligence-based automated tiering, where management of data in turn takes care of the storage bandwidths,” explains S Sridhar, Director, Enterprise Solutions Business, Dell.

EMC's Varma cites an example of a major telecom player in India and SAARC that recently benefited from the company's tiered storage offering. He says that previously the customer was manually selecting its non-performing or retention data and was placing it on cost-effective drives, like the SATA drive. “We helped the customer migrate the entire customer data from standard drives to tiered storage that involved a combination of flash drives, fiber channels and SATA drives,” he says. In this case, flash drives formed the highest performing tier, followed by the next tier of fiber channel, while SATA drives formed the lowest storage tier, which is a cost-effective solution for storing large volumes of inactive enterprise data. “Since Flash is a high performance but expensive storage technology, we did not want to overuse it and incur huge costs for the customer. So, we used just 9 percent flash drives in the overall storage of the enterprise, with 40 percent of the storage being fiber channel drive and 51 percent SATA drives. Within this storage architecture, we then introduced automated storage tiering that removed the manual intervention, improved response time of applications and enabled the company to save almost 40 percent upfront cost, which it would otherwise have to invest in various storage technologies,” adds Varma. STORAGE VISUALIZATION IDC defines storage visualization to include the ability to reallocate a heterogeneous collection of storage resources across storage systems and the capability to automate storage management functions. As a technology, storage virtualization has been around for quite some time now for the enterprises to evaluate and test. And it has been observed that enterprises are critically evaluating how storage virtualization technologies could be utilized to enable easy management of consolidated storage resources and ensure smooth data migration between different storage arrays. Venugopal of Hitachi says, “Apart from the regular business benefits of storage virtualization like common management of heterogeneous storage arrays and applying common processes to meet different SLAs of business applications, the most important business benefit for the customer is that they can now reuse their existing assets in the data center and ensure a far better return on assets.” He adds that companies that implement storage virtualization are able to reclaim about 65 percent of the capacity lying as toxic waste in their data center and that these customers can reduce their CAPEX by 3035 percent and OPEX by 40 percent. INNOVATIONS AROUND STORAGE VIRTUALIZATION Perceiving the fact that with the staggering data growth, ClOs would seriously look at implementing storage virtualization, technology vendors are honing their offerings in the space. Automated Storage Provisioning is a recent innovation that VMware has come out with to manage the storage needs within a virtual environment. One of the biggest challenges that existed in the virtual environment with respect to storage was that customers kept creating several hundreds and thousands of virtual machines, which in turn created huge sprawl. Nagarajan of VMWare tells us that generally their storage virtualization customers had to regularly maintain very complex Excel sheets containing notes like what are the different kinds of storage that the company has, which storage is meant for which application, which storage has how much free capacity, which Virtual Machines (VMs) are lying on which storage and so on. To address this, VMware came out with an offering that completely eliminated the manual intervention in managing and provisioning the virtual storage."In 2011, this issue was resolved when we came out with Vsphere 5.0, which came with automated storage provisioning under which administrators now have to

just create VMs and the rest is taken care by the hypervisor software, which would decide where the VM should run, at what point in time, in order to give desired performance. This has now drastically brought down the complexity of managing storage within virtual environments," Nagarajan adds. The concept of Virtual SAN is another important storage-linked innovation that VMware is set to introduce in its virtual environments in 2013. Minimum number of hard drives that a single server today comes with is two, with each hard disk of around 3 terabytes storage. In spite of the availability of so much free storage in number of servers that are spread across the enterprise data center, businesses today are not able to utilize this available storage for effective usage. Instead, they have to use an external shared common storage — the physical SAN — for enterprise use. Talking about how virtual SAN technology from VMware is set to change this situation in virtual environments, Nagarajan says, "With virtual SAN we could combine all the internal storage available on the servers and show this as an external shared storage for usage. As a result enterprises would be able to use the unutilized internal storage rather than paying for external storage arrays. This technology would be storage vendor agnostic and would be a part of the VMware hypervisor which would in turn enable managing the storage available on servers as simple as managing CPU or memory today." Converged infrastructure is another concept driven by virtualization that technology vendors like HP and Dell have started bringing into the market. The concept involves virtualizing all the heterogeneous resources like storage, server and networking creating an integrated virtualized resource pool that can be managed via a single console. This offering is specifically brought out for those industry verticals whose data storage and data management needs are huge and they at the same time need high performance of their business applications. “Today, responding to customer demands on time, managing IT budgets and performing with efficiency is the essence. And convergence in the data center is an emerging trend in IT that addresses the growing needs for agility, efficiency and quality to support the delivery of applications and IT services. Adoption of converged infrastructure lowers cost of running critical workloads, and enables faster infrastructure deployments and simplicity and speed of management," says Dell's Sridhar. The whole gamut of intelligent storage technologies is being today referred by a broad term called Software-Defined Storage (SDS). At this point in time, we can clearly see that storage vendors and virtualization vendors are working towards SDS, integrating software with the storage architecture in such a way that managing storage assets and smoothly scaling up and down according to changing needs of an enterprise becomes a reality. CLOUD STORAGE FOR DATA ARCHIVING AND BACKUP Companies have petabytes of data, which they can't afford to delete from a business perspective and they traditionally preserve it by spending on expensive on-premise data archiving and backup solutions. Using public cloud storage for data backup and archival is an interesting option today for the ClOs, as this can turn out to be much cheaper than the on-premise solutions. “From an IT perspective, by spending on data archiving and backup to be done on premise, enterprises are dealing with large costs in durably preserving this content indefinitely. This is one of the reasons that led us to invent and build Amazon Glacier, which is an extremely low-cost storage service that is optimized for data that is infrequently accessed and for which retrieval times of several hours are suitable," says Alyssa Henry, VP, Storage Services, Amazon Web Services. "In general, we're seeing a number of successful companies use AWS to both store their data and leverage it to develop greater insights. This is a trend we see across verticals, whether it's biotech, advertising, or e-commerce," he adds.

The company has also brought out several other cloud storage offerings that are customized for workloads that demand high performance. Software-defined storage is a clear technology trend that is emerging in the area of storage. Here technology vendors are coupling software and virtualization technologies along with storage so as to enable the enterprises to manage their storage environments in a much simpler way and at the same time improving the speed with which the enterprise applications could retrieve information from back-end storage. Though storage virtualization was not at the forefront for most enterprises, going forward, with the multi-fold increase in data— both structured and unstructured — enterprises are set to seriously evaluate this technology. Another interesting trend is how cloud storage is emerging as a cost-efficient yet effective option of storage for data backup and archival for the large enterprises. In an era where cost-cutting has become a mandate for the ClOs, they are sure to look at cloud storage for backup and archival as this is one of the ways in which they can shunt the money traditionally spent on storage of huge volumes of inactive data to other business critical operations. With an increase in storage and data management challenges in enterprises, both storage and virtualization technology Vendors are coming out with more storage-linked innovations to tap the growing market and enable the ClOs to consolidate their organization's storage environment in a way best suited to its particular storage and data needs. This is in fact, the right time for ClOs to analyze their storage architecture, identify the best suited storage technologies for the business and map out a strong storage strategy, keeping in mind the increasing volumes of enterprise data that it would have to keep pace with.