Critical Capabilities for Public Cloud Storage Services

Key Findings

• Public cloud storage is a viable alternative to building on-premises storage infrastructure for some uses. Several providers, though still evolving, offer highly competitive services with a range of capabilities, support and sales approaches.
• Cloud providers vary widely in the ability to meet enterprise IT requirements, and principally target developers and small or midsize businesses (SMBs).
• Cloud providers generally align around their own provider-specific interface APIs, or the OpenStack foundation API most new providers gravitate to.
• All providers focus on their capabilities in pricing, security and resiliency, but forward-looking vendors realize the importance of offering the reliable services, support, ease of management, turnkey capabilities and value-added services enterprises need.

Recommendations

• Assess the provider's vision, financial viability and track record to make sure the service and strategy closely align with your business and IT objectives. Cloud storage providers will become strategic partners, similar to large storage vendors.
• Select providers that provide a complete storage service solution that eliminates a do-it-yourself technology ensemble.
• Develop a two-year road map for these critical capabilities, and match them to the provider's current capabilities and future road map.

This document was revised on 10 January 2013. The document you are viewing is the corrected version. For more information, see the Corrections page on gartner.com.

Cloud storage has significant cost and agility benefits for organizations, but poses security, privacy and accessibility challenges. Organizations must carefully select a cloud storage provider through rigorous planning that involves a thorough understanding of the vendor's capabilities. A successful cloud storage infrastructure implementation depends on setting baseline requirements, establishing operational expectations and identifying business goals before launching a storage transformation project. Organizations must prepare for a long-term relationship with the cloud services provider, and should carefully evaluate and pilot the provider's services before placing critical data with the provider. In addition to the critical capabilities in this research, customers must clearly understand the long-term total cost of ownership (TCO) of cloud storage services by considering storage, bandwidth, data access, migration, internal and external support costs.

Introduction

This research focuses on cloud storage service providers, and does not include on-premises private cloud storage infrastructure vendors. Comparing cloud storage service providers requires a consideration of the technical capabilities, understanding the market the vendor intends to serve and assessing the provider's track record to address vertical industry (e.g., healthcare, financial, government, etc.) solutions. Understanding a provider's core capabilities is important, but so is understanding its value-added services and the road maps that build on the core capabilities. For organizations comfortable dealing with traditional storage vendors (e.g., EMC, NetApp and IBM), service and sales interaction expectations will require adjustments, as the organization moves from the technical and dedicated sales and support of a traditional vendor to the more generic sales and service approach taken by many cloud storage providers. Moving storage services into the realm of the cloud is not a fleeting activity, and requires a deep commitment to a cloud storage service provider due to the difficulty inherent in migrating data and services to an alternative provider. Organizations must choose wisely before proceeding with a cloud storage initiative.

Product Class Definition

Gartner research defines the public cloud storage services evaluated in this research as part of the software and hardware infrastructure as a service (IaaS) market. Vendors must supply their core storage services off-premises, and must include on-demand, elastic storage capabilities in a self-service model.

Critical Capabilities Definition

Cloud storage service offerings provide critical capabilities that define the product class and determine the suitability of the offering for customers. The capabilities of cloud storage service providers range from their technical proficiency in providing services to their pricing models, data protection and ability to offer valuable service and support. Vendors strive to excel in the areas they feel will attract customers in their targeted market segment, and may focus unevenly across the critical capabilities in this report. This results in differentiation across vendors, as they attempt to appeal to their targeted audience, workloads and use cases.

The critical capabilities considered for this research include:

• Accessibility: This is the ease with which customers can connect to the service, the performance of the service and the ability to securely read and write data. Options can include WAN optimization services and other on-premises integration services to enhance seamless data movement and access.
• Manageability: This refers to the ease of managing, monitoring and auditing the service through simple Web interfaces. Options can include integration with on-premises tools and specific change management features to simplify administration.
• Pricing: This comprises the simplicity and competitiveness of pricing, available payment options and assurance guidelines.
• Resiliency: This references system availability and the data protection mechanisms that guarantee robust uptime and disaster recovery. Additional options can include features such as write once, read many (WORM), versioning, incident analysis and integration with third-party independent software vendor (ISV) tools (such as backup, recovery and archiving).
• Security and Compliance: This refers to security controls, such as identity management, granular role-based access control, encryption mechanisms, anti-malware protection, forensics analysis and compliance controls related to physical security, industry and government regulations, and data eradication.
• Value-Added Services: These are the complementary services offered by providers, such as computing services, archiving services, hybrid cloud services, industry-specific offerings, custom SLAs, and consulting and support services.

Use Cases

This report considers the capability of public cloud storage providers to participate in four different use cases (see Table 1):

• Primary Storage: This is the delivery of data as part of ongoing operational activities, and as an extension to on-premises storage. Examples include employee file shares, software as a service (SaaS) application data and data to operate a virtualized server environment. Primary storage environments often include an intervening gateway appliance or file system software to accelerate performance, provide data manipulation and synchronize multiple users. Primary storage requires ease of access and management and data resilience similar in scope to on-premises data storage, but with improved economics.
• Backup: This is the protection of data similar to traditional on-premises backup environments, but using public cloud storage as the backup target. Backup environments may include agents running on desktops and servers, or backup servers/appliances that transmit backup data to the cloud directly or through an intervening gateway appliance. Backup requires high write and read performance, rapid recovery, regulatory compliance and resilience that should compare favorably with on-premises storage, but with better storage economics.
• Archive: The storage and infrequent access of data required for long-term retention and thus not often read. Archive environments require an application to manage content above and beyond what the cloud storage service provides, and often are provided as a solution that aggregates software, cloud storage and possibly intervening gateway appliances. Archive environments depend on data reliability/availability with economics that support long term storage and compliance with a wide range of regulations that may require regional placement of data.
• Content Distribution: The delivery of data for a widely dispersed set of consumers. Similar to primary data delivery, content distribution adds the dimension of regional specificity to content availability and requires synchronization to maintain consistent versioning over dispersed data centers. Content distribution, has similar requirements as primary data, with an emphasis on data movement and the availability of data centers or content distribution sites that are conveniently located to local users.

Source: Gartner (December 2012)

Inclusion Criteria

The inclusion criteria include storage service providers that offer core object or file-based storage services through a publicly defined API. Storage services are essential for file storage, backup data, archive or distribution environments, and often are associated with other application software and acceleration devices to create a complete solution. The vendor inclusion criteria qualify vendors that place a limited infrastructure at the customer location to enhance connectivity, security and performance.

To participate in this report, a vendor's services must meet the following minimum requirements:

• Offer APIs for data access over WANs or Internet connections, and protocols that include Internet APIs, such as REST, and/or network file access protocols, such as Network File System (NFS) and Common Internet File System (CIFS).
• Offer user-transparent and rapid storage capacity elasticity with on-demand accessibility. As customers store more data, cloud storage capacity must automatically expand to accommodate the data, albeit at a cost to the end consumer.
• Provide remote data storage for the target data in a robust facility with adequate infrastructure redundancy, and by adhering to various industry standards.
• Offer definable data security, reliability and availability as part of an SLA.
• Offer fault-resilient and secure network capabilities, including dedicated and shared links.
• Offer pay-as-you-go pricing for capacity and data transfer at a granular level (e.g., per gigabyte/month for capacity, and per gigabyte transfer/month for bandwidth).
• Have an established market presence, as demonstrated by the amount of petabytes under management, the number of clients or significant revenue attributable from storage service sales.

The cloud storage service market definition does not include storage for a cloud computing infrastructure. For example, the cloud storage vendors evaluated in this research are not required to offer storage connected to cloud computing instances. The control of those storage resources is through a cloud compute instance, and is not directly accessible as a stand-alone storage service. One example is Amazon Elastic Block Store (EBS), which is used for an Amazon Elastic Compute Cloud (Amazon EC2) computing instance. File sync and share (FSS), as well as turnkey backup and archive service providers, are not included in this report, since they offer fully integrated, turnkey solutions. However, FSS, backup or archive cloud providers may use the core storage services offered by the providers covered in this research.

Critical Capabilities Rating

Each product or service that meets our inclusion criteria has been evaluated on several critical capabilities (see Table 2 and Figure 1), on a scale from 1.0 (lowest ranking) to 5.0 (highest ranking).

Source: Gartner (December 2012)

Source: Gartner (December 2012)

To determine an overall score for each product in the use cases, the ratings in Table 2 are multiplied by the weightings shown in Table 1. These scores are shown in Table 3.

Source: Gartner (December 2012)

Product viability is distinct from the critical capability scores for each product. It is our assessment of the vendor's strategy, and of the vendor's ability to enhance and support a product throughout its expected life cycle; it is not an evaluation of the vendor as a whole. Four major areas are considered: strategy, support, execution and investment. Strategy includes how a vendor's strategy for a particular product fits in relation to the vendor's other product lines, market direction and overall business. Support includes the quality of technical and account support, as well as customer experiences with the product. Execution considers a vendor's structure and processes for sales, marketing, pricing and deal management. Investment considers the vendor's financial health, and the likelihood of the individual business unit responsible for a product to continue investing in it. Each product is rated on a five-point scale, from poor to outstanding, for each of the four areas, and is assigned an overall product viability rating.

Table 4 shows the product viability assessment.

Source: Gartner (December 2012)

The weighted capabilities scores for all use cases are displayed as components of the overall score.

Figure 2 shows the overall use case.

Source: Gartner (December 2012)

Figure 3 shows the content distribution use case.

Source: Gartner (December 2012)

Figure 4 shows the backup use case.

Source: Gartner (December 2012)

Figure 5 shows the archive use case.

Source: Gartner (December 2012)

Figure 6 shows the primary storage use case.

Source: Gartner (December 2012)

Vendors

Amazon S3

As an early and aggressive entrant into the cloud storage market, Amazon remains a leader in this space, and drives the offerings and behaviors of many competitors. Amazon's pricing remains the industry reference point, and with the introduction of Amazon Glacier, it provides attractive pricing even for basic archival storage. Amazon has had tremendous success in attracting ISVs that build solutions on top of its cloud offerings. Accessibility to Amazon S3 is provided through Amazon's unique API commands, with support for hybrid infrastructures delivered through the AWS Storage Gateway service or direct API integration. Although this approach is technically sound, enterprise customers will be left with the task of architecting, implementing and managing several elements of a cloud storage environment independently or via supported third-party solution, as opposed to purchasing a complete turnkey solution directly through Amazon. Service support remains focused on a broad set of customers, and although regional- and account-focused, it is evolving as Amazon addresses enterprise clients accustomed to working with large storage vendors for internal storage infrastructures. Amazon is developing its appeal to customers in various vertical industries by offering specific solutions for governmental agencies, and regulatory compliance coordination for the financial and healthcare industries. Further work is necessary (e.g., turnkey on-premises to cloud solutions), however, to fully integrate Amazon's cloud storage infrastructure with the on-premises storage of large enterprise IT organizations. Tighter integration with on-premises storage management tools is necessary to assist IT organizations looking to cloud storage to relieve capacity demands, and to provide more agility by moving data across private and public cloud infrastructures.

AT&T Synaptic Storage as a Service

AT&T Synaptic Storage as a Service has a focus on selling to midsize and enterprise-scale markets using EMC Atmos storage as the back-end infrastructure. AT&T claims high, double-digit growth rates from a base of several billion objects in its storage infrastructure. Atmos allows hybrid (private on-premises to public services) implementations to similar on-premises storage, and can be an advantage for customers already strategically aligned with EMC. The selling of AT&T Synaptic Storage as a Service occurs largely through existing AT&T services, the AT&T application service group and via the online cloud portal. Partnerships are maintained with, for example, cloud storage gateway vendors (such as Riverbed) to allow solution selling to provide customers with near-turnkey environments. Although currently limited to U.S. data centers, the storage as a service (STaaS) infrastructure is in place in EMEA, and AT&T plans expansion into Europe and beyond, as the customer base grows and customer workloads span regions. Thus, customers must explore the global data center distribution of the Synaptic service to verify its compatibility with regional content distribution requirements. Similar to other competitors, AT&T Synaptic Storage as a Service clients may proactively set data redundancy policies as they move data into the AT&T cloud storage services. Unlike other services, Synaptic allows users to enable data compression to save on data footprint size, and offers additional services, such as an integrated mobile FSS solution based on Oxygen Cloud. Clients that choose to use a private AT&T MPLS virtual private network (VPN) connection may obtain pricing advantages by avoiding ingress or egress charges.

Google Cloud Storage

Google Cloud Storage was launched in 2010, and targets the storage specific use cases in this research. It also is the underlying storage for other complementary products, such as Google App Engine, Google Compute Engine and Google BigQuery. Google Cloud Storage provides access through its proprietary RESTful API, but supports open authentication methods including OAuth 2.0. Google Cloud Storage is targeted primarily at developers, providing a simple and easy-to-access container to store objects. Google lacks significant sales/support focus to address the buying patterns and requirements of enterprise-class data centers, so cloud storage engagements with Google require a sophisticated customer willing to integrate the elements necessary to build a storage solution. Google Cloud Storage is compatible with the Amazon S3 API, which makes it easier for customers to move objects, and to attract developers who write applications compatible with S3. Unfortunately, a focus on turnkey solutions that includes on-premises hardware and software for larger clients is receiving little attention. Currently, customers can store data only in U.S. or European data centers. Google's biggest advantages are its potential to provide large-scale capacity for Google applications, and its business credibility as an experienced cloud provider for consumers and businesses.

HP Cloud Object Storage

HP made its public cloud services available via public beta in May 2012. HP's public cloud is built on OpenStack technology, and includes compute, storage and content delivery network (CDN) services. HP aims to differentiate its services from other OpenStack providers by building a rich ecosystem of ISVs, enterprise-class features and support, and by providing value-added capabilities to the OpenStack platform. HP offers 99.95% availability for its object storage services that is backed by free 24/7 support, including live phone and chat. Shortly after the launch of HP Cloud Object Storage, HP announced a CDN partnership with Akamai, and expanded the availability to an additional region (the U.S. East Coast) to complement its U.S. West Coast presence. Among OpenStack-based cloud storage providers, HP is well-positioned to understand enterprise IT storage requirements, due to its extensive hardware, software and service offerings. However, since HP Cloud Object Storage is new, HP must evolve and refine its architectural, geographical and service offerings.

IBM SmartCloud Enterprise Object Storage

IBM SmartCloud Enterprise (SCE) Object Storage, launched in 2010, is a key element of the SmartCloud Services family of cloud offerings that includes infrastructure, platform and application management services. The object storage offering leverages a Nirvanix-developed infrastructure that is dedicated to IBM services, and is housed in both IBM and Nirvanix data centers. IBM markets its offering for multiple use cases, such as backup and archive on its SCE offering utilizing the object storage capability. As the object storage product evolves, IBM intends to integrate it more tightly with other cloud computing and service offerings, such as IBM SCE+, SmartCloud Managed Backup and SmartCloud Archive. For example, the storage offering is not currently the back end for the backup or archive services, which use their own storage infrastructure and can create silos of capability. IBM's key advantages are its relationships and experience in selling to midsize and enterprise customers already comfortable with the current selling paradigms for on-premises storage hardware and software. SmartCloud Enterprise Object Storage is in an expansion phase in North America and Europe, with many proofs of concept (POCs) under way with clients using IBM's object storage for stand-alone public cloud storage, and as part of a larger IT project. Translating the POCs into production environments will be a key challenge for IBM, as it is for other cloud storage providers participating in the enterprise customer space. Given IBM's extensive on-premises storage portfolio, potential customers may benefit from the integration of on-premises and cloud capabilities as a single IBM solution, although this integration currently is achieved through the use of solutions from IBM business partners.

Internap AgileFILES

Internap is a publicly traded company (with revenue of $245 million in FY11) headquartered in Atlanta, Georgia. Internap started as a network service company, and in 2000 acquired CO Space to enter the co-location market. In 2011, the company launched an object storage service based on OpenStack Swift to serve business customers, becoming the first public cloud services provider to do so. In late 2011, Internap acquired Voxel, a provider of managed hosting and cloud services that complements Internap's cloud storage offering. Internap currently offers cloud services from its data centers in Amsterdam, Dallas, New York, Santa Clara, California, and Singapore, with plans to extend its footprint in 2013 to draw global clientele. Its cloud storage service is AgileFILES, and is differentiated through its network services, such as Managed Internet Route Optimizer (MIRO), a proprietary route optimization algorithm. However, lack of a major presence in the managed hosting and cloud computing IaaS segments is a severe limitation when selling cloud storage to enterprise customers.

Microsoft Windows Azure Blob Storage

Microsoft Windows Azure Blob Storage is growing rapidly, bolstered by the wide range and market penetration of Azure's other compute and application services. Object counts are in the trillions, growing at a rate of about 200% per year. This growth establishes Azure as one of the largest cloud storage vendors, along with Amazon. For storage workloads, Azure offers an array of storage services, such as object storage, durable NT File System (NTFS) volumes, Windows Azure Table storage and CDN. Geographic redundancy is possible over geographically separated (a minimum of 400 miles) data centers to provide disaster protection and distributed object delivery. Pricing continues to drop with a monetary commitment model that locks in pricing for large-scale customers. Microsoft is in a race to the bottom on pricing, and is currently reactive to competitive-pricing pressures from other cloud storage vendors. Playing catch-up to Amazon, Microsoft continues to offer attractive capabilities, but is in fast-follower mode on features and pricing, compared with Amazon, e.g., a low-cost alternative to Amazon's Glacier. Azure claims a transaction price advantage over much of its competition, but because pricing is rapidly changing, customers should validate any comparisons. The tiered set of support offerings is appealing across a wide range of customer types, with enterprise-class customers likely drawn to the more hands-on and fee-based premier tier of services. Microsoft is fostering a set of partnerships with cloud storage gateway vendors to offer a solution-oriented approach, and has an advantage in the wide reach and breadth of its sales and support workforce. A strong positive for Microsoft is its recent acquisition of StorSimple, a cloud storage gateway vendor.

Nirvanix Public Cloud Storage

Nirvanix is a privately held company that focuses exclusively on the cloud storage marketplace. Similar to other cloud storage vendors, Nirvanix claims very high growth rates, but uniquely, with a focus on enterprise-class storage customers. Nirvanix's sales and support are built on an engagement model familiar to large customers that routinely work with vendors such as IBM and EMC. This approach can be particularly attractive to premium customers interested, for example, in predictable all-inclusive monthly charges. However, smaller customers could find the pricing too high, compared with the a la carte pricing of cloud storage vendors that caters to a wider audience, such as for Amazon or Microsoft. Nirvanix offers several deployment models that range from the on-premises Nirvanix infrastructure to hybrid and off-premises public cloud storage. This capability can be appealing to large IT organizations that are testing the viability of cloud storage and wish to incrementally move into it. Unfortunately, Nirvanix does not offer a cloud compute capability for those who wish to extend their IT infrastructure further into the cloud than just data. This can hinder IT organizations that are planning a cloud road map that integrates compute instances, but could appeal to those keenly focused on data-oriented use cases and distribution.

Rackspace Cloud Files

Rackspace is a publicly traded company that offers a diverse suite of data center services, such as managed hosting, hybrid hosting, public and private cloud IaaS, cloud storage, hosted virtual desktops and SaaS. Along with NASA, Rackspace co-founded OpenStack, and has been a primary contributor to the Swift object store project under OpenStack. Swift is fast emerging as the storage foundation for many public cloud providers, such as HP, Internap and SoftLayer. Rackspace cloud storage (Cloud Files) is targeted primarily at developers with infrastructures located in Dallas, Chicago and London. Similar to other cloud storage vendors, Cloud Files integrates with Akamai's CDN, allowing delivery of static data (e.g., video, pictures, audio, Cascading Style Sheets [CSS], etc.) around the world. Through its pivotal role in founding and running OpenStack, Rackspace now offers simpler hybrid cloud environments for customers deploying OpenStack in the customers' private cloud. Rackspace also offers Cloud Block Storage to address customer demands for performance for file systems, databases and other input/output (I/O)-intensive applications. Rackspace has achieved impressive growth in revenue and customers for its cloud services in the past 12 months. Although the company prides itself on customer support, enterprise-scale customers expect deeper sales and consulting service engagements than Rackspace is currently delivering to help integrate cloud technologies with existing IT environments, and any consequent conversions and migration activities.

SoftLayer Object Storage

SoftLayer is a private hosting and cloud computing provider headquartered in Dallas. Its cloud services are branded as CloudLayer, and include computing, CDN and object storage services. Apart from its cloud services, SoftLayer also offers dedicated and managed hosting. Its cloud storage service is based on OpenStack Swift, introduced early in 2012, and targets SMB customers seeking integrated data center services. SoftLayer offers object storage service out of its data centers based in Dallas, Amsterdam and Singapore. To complement its object storage services, it also offers storage area network (SAN) as a service and electronic vaulting. SoftLayer has a limited market presence, compared with service providers such as Amazon, Nirvanix and AT&T. To attract enterprise customers, SoftLayer needs to add a strong professional service and support team, and to introduce more turnkey solution approaches for common use cases.