Virtualization Technology News and Information
VMblog's Expert Interviews: DataCore Software Talks Vision, New DataCore Programs, Accelerated Adoption and Expansion of Software-Defined Storage and Hyperconvergence


DataCore Software recently unveiled a series of new initiatives designed to simplify, automate and reward expansion of software-defined storage (SDS) and hyperconverged Virtual SAN throughout the globe.  To find out more, VMblog spoke with DataCore's chief product officer, Rizwan Pirani.  

VMblog:  Software-defined storage has been a rapidly growing market segment as users increasingly realize the benefits of a software vs. hardware approach.  What do you think are the factors behind this market shift?

Rizwan Pirani:  IT has been on a path to Software-Defined Infrastructure (SDI) for some time. Virtualization, cloud, and DevOps are all instances that are driven by or require SDI to work.

Compute virtualization has been a driver for IT efficiency for the past 15 years or so. Today, almost every workload is virtualized. Storage virtualization is lagging server virtualization because storage is far more complex and because data is the lifeblood of almost every company today. Compute can be ephemeral and can scale horizontally, while storage needs to be persistent, performant, and always available.

Software-defined storage (SDS) takes storage virtualization and adds additional services and value around it. The main value is in flexibility (avoiding hardware and vendor dependencies) and agility (simpler provisioning, management, and optimization).

Software-defined storage also offers many economic advantages such as:

  • The ability to integrate new technologies quickly
  • Extending the life of current storage hardware
  • Increasing the value of past and future investments
  • Hardware and vendor independence

Hardware refresh cycles represent one of the most challenging aspects for IT, especially storage hardware. Typically, an IT department will go through a storage refresh cycle every three to five years, but in some cases the hardware can be used for a longer period of time. Software-defined storage is very flexible and lets enterprises add new storage and technologies-whether it's AFAs, NVMe, containers or cloud storage-non-disruptively, so they can be ready to integrate whatever comes next onto the existing hardware.

Software-defined storage allows an organization to build that architecture. Once the latest and greatest hardware comes out, it can easily be integrated into the environment, helping to modernize the data center and increase infrastructure agility. With software-defined storage, there's no need to rip and replace when new technologies arrive.

The same is true about architectures. With SDS it is easier to adopt different types of storage (direct attach, arrays, JBODs, etc.) and adopt new configurations like hyperconverged or hybrid-converged with minimal disruption. This reinforces the idea of the flexibility and freedom offered by SDS.

VMblog:  How is the hyperconvergence market evolving over time?  Many of the hardware vendors are shifting from a hardware to a software offering, is this is sign of maturity or is the market shifting?

Pirani:  Not only are many of the traditional hardware vendors in the hyperconverged market shifting from a hardware to a software offering, the market as a whole has shifted from the earlier vision as "hyperconverged" primarily consisting of the convergence of compute, storage and network in one single hardware unit into one encompassing technology that is essentially more of a "hybrid-converged" infrastructure. This is a sign of the user's maturity in demanding hyperconverged systems that better meet their needs and vendors are creating optimized solutions that effectively match those requirements.  

Hybrid-converged is a superset of hyperconverged infrastructure that provides the same capabilities of compute, storage and networking in one "box," which is typically an x86 server loaded up with drives (either flash or HDD, or any combination thereof) but with the additional functionality of also connecting to storage that is external. In other words, it provides all of the benefits of hyperconverged infrastructure without sacrificing existing Storage Area Networks (SANs).

With a hybrid-converged infrastructure, users longer have to choose between a SAN or an HCI appliance-they can have them both. This way, the capacity of existing hyperconverged nodes can be expanded by connecting to an external SAN without needing to shut down the HCI 1.0 nodes, adding more storage or purchasing more nodes.

This is, again, an expected evolution towards the vision of software-defined infrastructure. The basic premise is that everything is defined, controlled, and managed via software, which means everything is flexible and there should be minimal hardware constrains. Hyperconverged is really a hardware-centric configuration model.

VMblog:  Many hardware storage vendors are adding more capable services to their arrays like mirroring and encryption.  What are the benefits of having this intelligence and the services in a software-defined layer across all storage versus buying it as part of a storage array?

Pirani:  There are a couple of advantages for IT organizations that decide to have an independent software-defined layer that offers services like mirroring or encryption. Here are a few:
  • The services and their management are consistent across all storage systems, independent of vendors or types of storage.
  • The services work across storage arrays. This means, for example, the system can dynamically auto-tier between an older array, cloud-based secondary storage, a newer flash array, and direct-attached NVMe storage, automatically.
  • A software-defined storage layer eliminates migrations. Decommissioning, or adding a new array is a simple as adding it to the pool and assigning a performance tier level.
  • IT avoids being locked-in by a storage vendor on multi-year contracts for their technology. Instead they have the freedom to use any storage system they want, including low-cost commodity storage, which can perform great given the services are in a separate layer.
  • SDS software is usually sold under a perpetual license, which can be very advantageous financially. The software can be upgraded easily. In DataCore's case, we have moved away from charging for individual features, which means every customer (including those who purchased DataCore years ago) enjoys all of the features in their license tier, including auto-tiering, thin provisioning, mirroring, and continuous data protection.

VMblog:  What should users look for when choosing the optimal software-defined storage solutions?

Pirani:  Buyers should look at three main elements when evaluating software-defined storage: 
  1. Flexibility. This is one of the main reasons why IT departments should be building their strategy on top of software-defined infrastructure. Therefore, it is important that the platform allows full flexibility of hardware choices, network (iSCSI or FibreChannel), API accessibility, container-readiness, etc.
  2. Completeness. An IT department requires a solution that provides all the services that make a powerful SDS deployment: dynamic auto-tiering, performance acceleration, mirroring and snapshots, continuous data protection, thin provisioning, and ease of management.
  3. Maturity. Today, data is the lifeblood of a company. Your data is your business. It is important to be able to trust a software platform that is reliable, proven, and tested.

VMblog:  What other market trends do you find interesting in the space and how will they impact DataCore's growth?

Pirani:   I find the market trends of data analytics and cloud maturity particularly interesting. In terms of data analytics, as more enterprises look toward big data for future opportunities, it can become the cornerstone of positive transformation when collectively transitioned into actionable insights. We will increasingly employ the use of algorithms, machine leaning, and artificial intelligence (AI) to derive insights. With the collection, synthesis, analytics and visualization models now available, big data is being transformed into data-driven intelligence and this will be a trend that continues to impact the industry in the coming years.

The evolution, however, has only just begun. When data analytics meets "predictability," the immense power of "prevention is better than cure" is possible. The leverage of predictability relieves invaluable human cycles for more abstracted, sophisticated transactions while allowing compute/machine learning/AI to forecast remedies pre-emptively.

These logical deductions performed pre-emptively create a data-driven set of remedies. As businesses look toward big data for future opportunities and that data continues to get bigger and bigger, IT departments will need to embrace technologies such as software-defined storage to confront these challenges and achieve the promise that big data holds.

Regarding cloud maturity, I believe the IT industry has finally realized the place and time for cloud. It is not the panacea everyone expected a few years ago. It is also not a cost-cutting tool. The industry now understands the cloud is one more tool at its disposal, and that datacenters are going to be around for a while.

By the way, the line between ‘traditional' datacenters and the cloud has been blurring, as the former are now mostly virtualized resources on a co-location environment. The evolution towards software-defined infrastructure, whether those look like virtualized resources or public cloud, is what will give IT the flexibility and freedom.

Most IT departments should know by now what they plan to host in the cloud and what they plan to host themselves. There are many reasons to do one or the other. The reality is that the future is hybrid, IT needs to build infrastructure that leverages the cloud as an extension of the on-premises infrastructure.

The first step is to become software-defined. To break the silos, achieve vendor independence, and break free from three-year vendor-imposed refresh cycles.

I'll say one last thing about trends: remember that they come and go. Every trend has a new set of cool vendors. There is a group of storage array vendors and there are the AFA and HCI vendors. The risk is investing with one of these vendors for the long term, ignoring the fact that technology comes in cycles and that in a few years there are probably going to be a new set of technologies with a new set of vendors you want to do business with. This is one more reason to go software-defined; you can have the freedom to use whatever technology and whichever vendor your organization needs at any time.

VMblog:  How does DataCore enable customers to shift storage spending from a passive, recurring expense to a more valuable investment in software-defined infrastructure?

Pirani:  We mentioned the ability of SDS to provide advanced data services across old and new storage systems, but there are three important financial benefits IT departments that deploy DataCore enjoy: 

First, is Parallel I/O technology and advanced caching. I realize that most companies promise high performance, but there is only one company that smashed the previous record based on the SPC benchmark. Most DataCore customers see a 5X increase in performance through a variety of technologies. This means an older storage array can produce flash-like performance, which could delay or avoid a storage refresh cycle.

Second, thin provisioning and intelligent pooling result in massively increased storage utilization. Some IT departments see up to 50% more storage efficiency, which again can translate into significant savings in storage hardware costs.

Third is efficiency. Every IT department is under-staffed. Storage teams spend way too much time managing storage, provisioning for users, and fork-lift migrating to new arrays. All this work is essentially eliminated with an intelligent SDS layer that moves around data as needed, provisions storage from the same VMware and Hyper-V consoles where virtual servers are created, and provides a set of services that automate storage management.

VMblog:  How does DataCore help users to maximize IT infrastructure performance, availability and utilization?

Pirani:  To get better performance and availability, enterprises can't just "rip and replace" hardware every few years because it's both expensive and complicated. DataCore helps to optimize the underlying infrastructure, increasing the performance, availability and utilization of all applications.

DataCore's software works with existing infrastructure, extending the useful life of the equipment while giving companies the flexibility to pick the hardware, technologies and platforms to meet their needs and leverage existing IT investments when possible.

Better utilization of current infrastructure and seamless integration of advanced technology reduces costs and gives users more control on when to make hardware investments. DataCore's services are also hardware-agnostic, providing a unified storage layer for applications and management across different technologies and infrastructure vendors, easing administrative burdens. New technologies that work with standard x86 hardware can be easily integrated, from AFAs to NVMe, supplementing the capabilities of the environment and simplifying the process of assigning the right storage tier to each application.

To answer your question more specifically, DataCore developed a series of very unique technologies, including Parallel I/O and advanced caching, that eliminate bottlenecks and optimize all storage systems, old and new. The average DataCore customer sees a 5x performance increase. We have been holding the SPC price-performance benchmark for years. This means that old arrays may be able to run as fast as some of the newer AFAs. Further, dynamic auto-tiering across storage systems allows an IT department to add a little direct attached NVMe and enjoy incredible performance gains.

DataCore also has been a leader in terms of availability, supporting synchronous mirroring in local and metro clusters, asynchronous replication for disaster recovery, and continuous data protection, which is like a time machine to undo any damage from ransomware attacks. The most important aspect to ensure availability is the recovery-recovery is instantaneous and automatic, with zero touch fall back and re-build. Our technology is mature, and it just works. Many of our customers have seen zero downtime for many years.

In terms of utilization, we talked about the ability to get maximum value from existing investments. There is also thin provisioning, another technology we developed and patented. It simply eliminates the inefficiencies of storage allocations. Combined with auto-tiering across dissimilar storage systems, organizations can see 50% to 70% increase in storage utilization.

VMblog:  DataCore recently announced a new series of initiatives designed to simplify, automate and reward expansion of its software-defined storage (SDS) and hyperconverged Virtual SAN products (e.g. product enhancements, a simplified licensing/pricing model, new customer loyalty program and community platform).  As one of the team members directly involved in developing/overseeing these programs, can you tell me a little bit more each and what this means to your customers?  Your partners?  The market as a whole?

Pirani:  Partners will enjoy community forums where they (and their customers) can interact with other DataCore users to exchange best practices or discuss important topics.  However, the main value for partners is the new simplified pricing. It is based in two factors: the deployment model and the amount of capacity (TBs). The new pricing model includes all of the features and has no limitation in terms of number of nodes, so an IT department has the flexibility to add nodes as needed.

The new initiatives ensure that customers derive higher value from existing hardware and software investments, driving accelerated adoption and expansion of software-defined storage technologies worldwide. Elements such as stepped-up discounts based on lifetime capacity consumption as well as no-charge feature upgrades will make it even more attractive for the market to scale up and out for added resiliency and throughput in pursuit of real-time, always-on data.

DataCore’s new simplified licensing/pricing model also offers greater flexibility, allowing customers to transition at their own pace from conventional 3-tier SANs to 2-tier servers SANs, or quickly collapse into hyperconverged systems – whether on-premises, at the edge or in the cloud.

As IT department are looking to reap the benefits of the software-defined datacenter, customers will be able to quickly realize the performance, uptime and flexibility advantages of software-defined storage. This will help them spend less time on repetitive tasks and expand the technology to cover more of their IT footprint, including additional workloads or datacenters.  

VMblog:  Is DataCore doing anything in the space of predictive analytics and machine learning?

Pirani:  Predictive data analytics is a domain that DataCore is uniquely suited to helping its customers capitalize on. In reality, DataCore has been using machine learning for many years. For example, our patented caching algorithms use machine learning to provide the most efficient caching possible.

Today, DataCore servers can transmit high-level telemetry data to a central server. This data (which is anonymous and opt-in) is available now for support purposes. We are currently working on innovative ways to leverage it to analyze it using a series of AI-assisted analysis to do predictive support, analyze patterns, or provide other intelligent services to customers.

VMblog:  DataCore is expanding its support for critical new technologies such as containers and NVMe.  Can you tell me more about these areas?

Pirani:  Container data services are expected to revolutionize the computing industry with innovative solutions, and we are well placed to offer monumental value in this space. It is very clear that containers are here to stay and will play a big role in IT in the coming years - maybe even over decades.

Containers bring unprecedented mobility, ease and efficiency for rapidly deploying and updating applications. At the same time, they introduce new storage challenges-creating an opportunity for innovation and differentiation.

As container deployments move from evaluation and testing phases to production deployments, IT organization require an ability to deliver the same data storage services that are currently providing to monolithic application architectures. More importantly, a solution has to be capable of providing shared storage to existing virtualized and bare-metal application infrastructures, as well as allow DevOps engineers to consume storage on-demand, ensure stateful application data is persisted, and provide the same level of availability and performance as currently provided to the traditional application infrastructures.

Software-defined storage can provide an ability for administrators to present persistent storage to container hosts deployed as VMs on virtual hosts, with the ability to provide persistent storage to container hosts deployed on bare-metal as a next step. The presentation of the persistent storage should be done through native controls of orchestration solutions, like Kubernetes, and leverage advanced storage capabilities like CDP, auto-tiering and synchronous mirroring.

As a result, users can manage the provisioning of storage to container deployments, with the same platform as the rest of the application workloads, and provide the same level of enterprise storage services required for all critical production environments.

VMblog:  What is your ultimate vision for DataCore's product lines/technology?

Pirani:  DataCore continues to be a technological leader in the software-defined storage and hyperconverged infrastructure market. Its products are immensely applicable to enterprises today, as evidenced by the exceptional growth of the company and diversity of the partner community. The deep emphasis on features such as high availability and disaster recovery, along with the plethora of data services at exceptional performance levels, has been quoted by many customers, partners-- and even competitors. But perhaps the biggest endorsement of value stems from the ROI achieved by customers by the elimination of the incessant hardware refresh costs in their datacenters.

With this stable and enormous base of exceptional technical value, we are now primed to transform datacenters with focused investments in modern technologies. The applicability of prescriptive services along with proactive support for an efficient datacenter will help make DataCore the secret weapon for smarter IT departments that will use SDS as the foundation for modernizing their datacenter and delivering business value.


Published Thursday, November 08, 2018 10:48 AM by David Marshall
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
<November 2018>