Formerly known as Wikibon
Search
Close this search box.

Pure Extends Hybrid Cloud Data Plane Capabilities

Premise

The Pure Storage recent acquisition of file system specialist Compuverde is a tuck-in acquisition that extends Pure’s reach and capabilities initially into the mid-tier file markets. In addition, Pure intends to leverage Compuverde’s technology to provide software defined solutions to meet requirements for hybrid cloud file system customers. This is a low risk TAM expansion play targeting markets currently dominated by NetApp and Dell Technologies.

Why Flash is Accelerating Hybrid Cloud Adoption

Flash storage is one factor accelerating the move to hybrid cloud. Traditional magnetic storage has access times measured in milliseconds. Reading data held in memory 200 miles away in a cloud would take the same time as going to a local magnetic disk. However, with the speed of data in flash dropping to a low of 50 microseconds (about 500 times faster than magnetic) and continuing to drop, the datacenter has to be less than 2,000 feet away to achieve the same result.

Hybrid applications solve this speed of light problem by moving code and services to the data where possible. This enables these  applications to operate in parallel on many 100s and soon 1,000s of times more data than traditional systems. This approach is significantly more complex than a monolithic public cloud. However, the hybrid cloud approach is essential to enable data driven applications in real-time or near real-time.

In addition to the speed of light problem above, hybrid cloud also lowers the cost of moving data, allows data to remain where it is created, and simplifies meeting regulatory and business risk  requirements (see “Who’s Data is it?” section below).

Another key capability of flash is high bandwidth, which allows the same physical data to be shared across multiple virtual copies or versions of the data. Hybrid cloud storage can use metadata to manage the copies, can save time and money by avoiding moving data, and save money by sharing the same physical flash across multiple virtual copies. The copies can also be shared across nodes at distance with minimal data being shipped to keep the copies synchronized with each other.

Wikibon believes it is an imperative for storage vendors to embrace this complexity and develop strong distributed data services to support all hybrid clouds topologies for block,  file and object data.

Pure Acquires Compuverde

Pure is acquiring startup Compuverde, a Swedish file system software  company. Compuverde focuses on very large file systems managed across multiple locations. It is a software solution running on commodity servers. This is suitable for offering file services across a hybrid cloud data plane. This is also a solution for IaaS cloud providers and SaaS providers.

Compuverde has a broad set of partnerships, and for example supplies the software behind IBM Spectrum NAS. Pure will continue to support Spectrum NAS and expand the relationship with IBM.

Pure also recently acquired StorReduce for its hybrid cloud data services. Pure is using this technology to enable holding a local cache of recovery file data on high-performance Pure FlashBlade storage. This on-premises data is automatically synchronized with low-cost file storage in a public cloud.    

Who’s data is it, anyway?

One of the growing challenges for enterprises in a cloud world is clear ownership and privacy of data. It can be an amusement when an individual Googles about Tesla cars, and then receives an avalanche of emails advertising cars. It is more troubling if a SaaS vendor has access to all the sales calls made to companies, and sells that information to a third party to answer the question “who is the best person to call at this company”. It is also troubling if a cloud vendor monitors a SaaS vendor’s success to determine if they should launch competitive software. It is very troubling if a cloud provider complies with a court subpoena for your enterprise data without your knowledge.

A general solution to this problem uses hybrid cloud technology to hold important data directly under enterprise control and to hold less important data at the lowest cost in a cloud. A specific solution is to move services to execute where the data is, rather than moving the data to where the services are. The result is a more secure and cheaper way of storing and executing code against the data. The data is then available faster to access (no delay in moving it), which allows more value to be extracted from the data with data-driven hybrid applications.

The Complexity of Hybrid Cloud

For most enterprises, hybrid clouds are already a reality. Wikibon has introduced a Hybrid Cloud Taxonomy, which is illustrated in Figure 1 below.

 

Hybrid Cloud Taxonomy
Figure 1 – Wikibon Hybrid Cloud Taxonomy
Source: © Wikibon 2019.

Figure 1 – Wikibon Hybrid Cloud Taxonomy Source: © Wikibon, 2019

The degree of state, integration, automation and support for hybrid applications increase from left to right. The planes at the bottom of Figure 1 run in software across all nodes in a hybrid cloud, and need to be more sophisticated as you traverse from left to right. The planes include network plane(s), data planes(s), control planes, and security & compliance planes. All the planes need to be software defined and support a distributed model across all the nodes of a hybrid cloud.

Focusing on data, the integrated data plane is a key foundation of hybrid cloud. It provides data orchestration and automation services and connects closely via APIs with the network, control, security & compliance planes.

Hybrid cloud data planes are distributed and are more complex than traditional array-based data services. These distributed services need to be supplied as software defined services running on commodity hardware. The need to provide software for different performance characteristics, from pools of low-latency NVMe drives on high-speed networks to lower performance lower-cost alternatives. They need common management frameworks and connect to overall control planes.

So, if this is more complex than a simple monolithic cloud, why bother? This new hybrid cloud environment will begin by supporting traditional applications, and will see some benefit. However, over time data-driven hybrid applications will appear as developers, SaaS vendors, and business executives start to understand how hybrid applications can exploit data in parallel across the hybrid cloud. These solutions can utilize far greater amounts of data as soon as the data is available from any source in any location, including on-premises, the Edge, and SaaS and other clouds.

The results from the hybrid applications can be fed back in real-time into existing systems of record to help automate and accelerate business processes. An example would be monitoring all the very large number of factors from different sites that lead to demand for a product, and being able to adjust prices automatically in real-time or near real-time to optimize revenue and customer satisfaction.

The hybrid applications will drive requirements for more advanced services from distributed file systems. These services include distributed orchestration,  automation, distributed protection, and data distribution against disasters, cyber attacks and ransomware. One data service example is knowledge of the data latency and bandwidth from any node to any other node, which will allow optimal placement and orchestration of data seamlessly for hybrid applications. Other data plane services include end-to-end encryption, hybrid monitoring services against performance and availability SLAs, deep integration with advanced network plane capabilities, and many more.  

Pure’s Hybrid Cloud Data Plane Services

Pure Storage is a leading flash storage vendor. Pure Storage has embraced a hybrid cloud vision, and is investing in hybrid cloud services within an integrated data plane. With the Compuverde acquisition, Pure is extending that vision to file storage.

The Pure hybrid cloud data plane file services are deduplication, compression, thin provisioning, encryption, and data protection. Compuverde brings extended NFS & SMB protocol options, snapshots, and replication services. Pure also offers cloud API plug-ins and orchestration for VMware and AWS, as well as status monitoring and predictive support with the Pure1 SaaS offering.

These services are good, but table stakes. Pure will need to offer features such as distributed global namespaces, and the ability to deal with very large files and very small files, and end-trend support for hybrid applications. Pure will also need to provide and support APIs and distributed data services that enable deep integration into network planes, control planes, as well as connections with other data planes.

Hybrid Cloud File Alternatives

Dell & NetApp are the major players in file storage for on-premises equipment. IBM is taking an aggressive multi and hybrid cloud approach with its Spectrum series of software storage solutions.

Dell Technologies offers Isilon, a distributed file system for NFS & CIFS, and Elastic Cloud Storage (ECS) object storage. Both Isilon and ECS are available as software defined storage on commodity hardware, an important step in supporting hybrid cloud. Dell’s emphasis is on providing flexible consumption models for on-premises private clouds with mainly backup to public clouds.

IBM Spectrum NAS is a software defined storage package, which uses Compuverde technology. Pure will continue to support Spectrum NAS and expand the relationship with IBM. All the Spectrum storage software is software defined and supports IBM and other storage systems. IBM Spectrum Virtualize and Spectrum Virtualize for Public Cloud provide hybrid and multi-cloud supported on IBM Cloud and on AWS.

NetApp hybrid cloud solution is Cloud Volumes ONTAP, which provides NFS & CIFS cloud native file services for data running though AWS EC2 instances to NetApp ONTAP arrays in the same collocation (usually Equinix) as AWS, with a low-latency Direct Connect Gateway to AWS. The equivalent service is available with Microsoft Azure. NetApp also offers Cloud Manager as a hybrid cloud control plane. NetApp has good file mindshare with excellent file system capabilities, and a software defined solution running on native AWS and Azure. The hybrid cloud data services are limited at the moment to general control from Cloud Manager.

Conclusions  

Pure Storage has embraced a hybrid cloud vision, and is acquiring and investing in hybrid cloud services for an integrated data plane. At the moment IBM & Pure Storage are leading in software defined storage for hybrid cloud, albeit early in the first innings. Wikibon believes that all storage vendors will need to embrace a similar strategy, and develop strong hybrid cloud data services for both block, file and object data. This enables easier development and deployment of hybrid applications with the business improved business costs and revenue. There is plenty of white space for innovation available, and plenty of opportunities for integrating with partners and other storage vendors.

Action Item

A key element of enterprise digital transformation is enabling access to data and extracting value from data as quickly as possible, wherever it is situated. Hybrid applications are a key technology to utilize these data assets to enable automation and acceleration of business processes, with direct impacts on cost and revenue.

These hybrid applications need rich hybrid cloud topologies and advanced services. Establishing a high-function hybrid cloud environment to support these applications is a strategic imperative for all enterprises. This should be understood and supported by all levels of executive leadership in IT, within the lines of business, and at board level.  

Article Categories
You may also be interested in

Book A Briefing

Fill out the form , and our team will be in touch shortly.
Skip to content