While a few well-publicized web 2.0 companies are taking great advantage of foundational big data solution that they have themselves created (e.g. Hadoop), most traditional enterprise IT shops are still thinking about how to practically deploy their first business-impacting big data applications – or have dived in and are now struggling mightily to effectively manage a large Hadoop cluster in the middle of their production data center. This has led to the common perception that realistic big data business value may yet be just out of reach for most organizations – especially those that need to run lean and mean on both staffing and resources.
This new big data ecosystem consists of scale-out platforms, cutting-edge open source solutions, and massive storage that is inherently difficult for traditional IT shops to optimally manage in production – especially with still evolving ecosystem management capabilities. In addition, most organizations need to run large clusters supporting multiple users and applications to control both capital and operational costs. Yet there are no native ways to guarantee, control, or even gain visibility into workload-level performance within Hadoop. Even if there wasn’t a real high-end skills and deep expertise gap for most, there still isn’t any practical way that additional experts could tweak and tune mixed Hadoop workload environments to meet production performance SLA’s.
At the same time, the competitive game of mining of value from big data has moved from day-long batch ELT/ETL jobs feeding downstream BI systems, to more user interactive queries and business process “real time” applications. Live performance matters as much now in big data as it does in any other data center solution. Ensuring multi-tenant workload performance within Hadoop is why Pepperdata, a cluster performance optimization solution, is critical to the success of enterprise big data initiatives.
In this report we’ll look deeper into today’s Hadoop deployment challenges and learn how performance optimization capabilities are not only necessary for big data success in enterprise production environments, but can open up new opportunities to mine additional business value. We’ll look at Pepperdata’s unique performance solution that enables successful Hadoop adoption for the common enterprise. We’ll also examine how it inherently provides deep visibility and reporting into who is doing what/when for troubleshooting, chargeback and other management needs. Because Pepperdata’s function is essential and unique, not to mention its compelling net value, it should be a checklist item in any data center Hadoop implementation.
To read this full report please click here.
Virtualization has matured and become widely adopted in the enterprise market. HyperConverged Infrastructure (HCI), with virtualization at its core, is taking the market by storm, enabling virtualization for businesses of all sizes. The success of these technologies has been driven by an insatiable desire to make IT simpler, faster, and more efficient. IT can no longer afford the time and effort required to create custom infrastructure from best-of-breed DIY components.
With HCI, the traditional three-tier architecture has been collapsed into a single system that is purpose-built for virtualization. In these solutions, the hypervisor, compute, storage, and advanced data services are integrated into an x86 industry-standard building block. The immense success of this approach has led to increased competition in this space and the customers are required to sort through the various offerings, analyzing key attributes to determine which are significant.
One of these competing vendors, Pivot3, was founded in 2002 and has been in the HCI market since 2008, well before the term HyperConverged was used. For many years, Pivot3’s vSTAC architecture has provided the most efficient scale-out Software-Defined Storage (SDS) system available on the market. This efficiency is attributed to three design innovations. The first is their extremely efficient and reliable erasure coding technology called Scalar Erasure Coding. Conversely, many leading HCI implementations use replication-based redundancy techniques which are heavy on storage capacity utilization. Scalar Erasure Coding from Pivot3 can deliver significant capacity savings depending on the level of drive protection selected. The second innovation is Pivot3’s Global Hyperconvergence which creates a cross-cluster virtual SAN, the HyperSAN: in case of appliance failure, a VM migrates to another node and continues operations without the need to divert compute power to copy data over to that node. The third innovation has been a reduction in CPU overhead needed to implement the SDS features and other VM centric management tasks. Implementation of the HCI software uses the same CPU complex as business applications, this additional usage is referred to as the HCI overhead tax. HCI overhead tax is important since the licensing cost for many applications and infrastructure software are based on a per CPU basis. Even with today’s ever- increasing cores per CPU there still can be significant cost saving by keeping the HCI overhead tax low.
The Pivot3 family of HCI products delivering high data efficiency with a very low overhead are an ideal solution for storage-centric business workload environments where storage costs and reliability are critical success factors. One example of this is a VDI implementation where cost per seat determines success. Other examples would be capacity-centric workloads such as big data or video surveillance that could benefit from a Pivot3 HCI approach with leading storage capacity and reliability. In this paper we compare Pivot3 with other leading HCI architectures. We utilized data extracted from the alternative HCI vendor’s reference architectures for VDI implementations. Using real world examples, we have demonstrated that with other solutions, users must purchase up to 136% more raw storage capacity and up to 59% more total CPU cores than are required when using equivalent Pivot3 products. These impressive results can lead to significant costs savings.
Nutanix XCP For Demanding Enterprise Workloads: Making Infrastructure Invisible for Tier-1 Ent. Apps
Virtualization has matured and become widely adopted in the enterprise market. Approximately three in every five physical servers are deployed in a virtualized environment. After two waves of virtualization, it is safe to assume that a high percentage of business applications are running in virtualized environments. The applications last to deploy into the virtualized environment were considered the tier-1 apps. Examples of these include CRM and ERP environments running SAP NetWeaver, Oracle database and applications, and Microsoft SQL Server. In many 24X7 service industries, Microsoft Exchange and SharePoint are also considered tier-1 applications.
The initial approach to building virtualized environments that can handle these tier-1 applications was to build highly tuned infrastructure using best of breed three-tier architectures where compute, storage and networking were selected and customized for each type of workload. Newer shared storage systems have increasingly adopted virtualized all flash and hybrid architectures, which has allowed organizations to mix a few tier-1 workloads within the same traditional infrastructure and still meet stringent SLA requirements.
Enter now the new product category of enterprise-capable HyperConverged Infrastructure (HCI). With HCI, the traditional three-tier architecture has been collapsed into a single software-based system that is purpose-built for virtualization. In these solutions, the hypervisor, compute, storage, and advanced data services are integrated into an x86 industry-standard building block. These modern scale-out hyperconverged systems combine a flash-first software-defined storage architecture with VM-centric ease-of-use that far exceeds any three-tier approach on the market today. These attributes have made HCI very popular and one of the fastest growing product segments in the market today.
HCI products have been very popular with medium sized companies and specific workloads such as VDI or test and development. After a few years of hardening and maturity, are these products ready to tackle enterprise tier-1 applications? In this paper we will take a closer look at Nutanix Xtreme Computing Platform (XCP) and explore how its capabilities stack up to tier-1 application workload requirements.
Nutanix was a pioneer in HCI and is widely considered the market and visionary leader of this rapidly growing segment. Nutanix has recently announced the next step - a vision of the product beyond HCI. With this concept they plan to make the entire virtualized infrastructure invisible to IT consumers. This will encompass all three of the popular hypervisors: VMware, Hyper-V and their own Acropolis Hypervisor. Nutanix has enabled app mobility between different hypervisors, a unique concept across the converged system and HCI alike. This Solution Profile will focus on the Nutanix XCP platform and key capabilities that make it suitable for teir-1 enterprise applications. With the most recent release, we have found compelling features appropriate for most tier-1 application workloads. Combined with the value proposition of web-scale modular architecture this provides an easy pathway to data-center transformation that businesses of all sizes should take advantage of.
Storage should be the most reliable thing in the data center, not the least. What data centers today need is enterprise storage that affordably delivers at least 7-9's of reliability, at scale. That's a goal of less than three seconds of anticipated unavailability per year - less than the reliability of most data centers.
Data availability is the key attribute enterprises need most to maximize their enterprise storage value, especially as data volumes grow into scales. Yet traditional enterprise storage solutions aren’t keeping pace with the growing need for greater than the oft-touted 5-9’s of storage reliability, instead deferring to layered on methods like additional replication copies, that can drive up latency and cost, or settling for cold tiering which zaps performance and reduces accessibility.
Within the array, as stored data volumes ramp up and disk capacities increase, RAID and related volume/LUN schemes begin to fall down due to longer and longer disk rebuild times that create large windows of vulnerability to unrecoverable data loss. Other vulnerabilities can arise from poor (or at best, default) array designs, software issues, and well-intentioned but often fatal human management and administration. Any new storage solution has to address all of these potential vulnerabilities.
In this report we will look at what we mean by 7-9’s exactly, and what’s really needed to provide 7-9’s of availability for storage. We’ll then examine how Infinidat in particular is delivering on that demanding requirement for those enterprises that require cost-effective enterprise storage at scale.
All of the trends leading towards the world-wide Internet of Things (IoT) – ubiquitous, embedded computing, mobile, organically distributed nodes, and far-flung networks tying them together - are also coming in full force into the IT data center. These solutions are taking the form of converged and hyperconverged modules of IT infrastructure. Organizations adopting such solutions gain from a simpler building-block way to architect and deploy IT, and forward-thinking vendors now have a unique opportunity to profit from subscription services that while delivering superior customer insight and support, also help build a trusted advisor relationship that promises an ongoing “win-win” scenario for both the client and the vendor.
There are many direct (e.g. revenue impacting) and indirect (e.g. customer satisfaction) benefits we mention in this report, but the key enabler to this opportunity is in establishing an IoT scale data analysis capability. Specifically, by approaching converged and hyperconverged solutions as an IoT “appliance”, and harvesting low-level component data on utilization, health, configuration, performance, availability, faults, and other end point metrics across the full worldwide customer base deployment of appliances, an IoT vendor can then analyze the resulting stream of data with great profit for both the vendor and each individual client. Top-notch analytics can feed support, drive product management, assure sales/account control, inform marketing, and even provide a revenue opportunity directly (e.g. offering a gold level of service to the end customer).
An IoT data stream from a large pool of appliances is almost literally the definition of “big data” – non-stop machine data at large scale with tremendous variety (even within a single converged solution stack) – and operating and maintaining such a big data solution requires a significant amount of data wrangling, data science and ongoing maintenance to stay current. Unfortunately this means IT vendors looking to position IoT oriented solutions may have to invest a large amount of cash, staff and resources into building out and supporting such analytics. For many vendors, especially those with a varied or complex convergence solution portfolio or established as a channel partner building them from third-party reference architectures, these big data costs can be prohibitive. However, failing to provide these services may result in large friction selling and supporting converged solutions to clients now expecting to manage IT infrastructure as appliances.
In this report, we’ll look at the convergence and hyperconvergence appliance trend, and the increasing customer expectations for such solutions. In particular we’ll see how IT appliances in the market need to be treated as complete, commoditized products as ubiquitous and with the same end user expectations as emerging household IoT solutions. In this context, we’ll look at Glassbeam’s unique B2B SaaS SCALAR that converged and hyperconverged IT appliance vendors can immediately adopt to provide an IoT machine data analytic solution. We’ll see how Glassbeam can help differentiate amongst competing solutions, build a trusted client relationship, better manage and support clients, and even provide additional direct revenue opportunities.
End-user mobility and the fast growth of data are increasingly pushing file storage and sharing into the cloud. Sharing data through an easy-to-use cloud service keeps globe-hopping users happy and productive, while technologies like cloud storage gateways enable IT to govern across on-premise and cloud storage as a single infrastructure. However, the big goals of these two groups often crash like fast ships on ramming speed. What end-users want as consumers is fast, easy, and always-on while IT requires fully secure, controlled, and ultimately cost-effective. This imbalance creates challenges: concentrate on end-user usability and governance suffers; concentrate on governance and usability diminishes.
How can live files and business data move with users, and IT control and governance follow the data wherever it goes? Globally mobile end-users need to efficiently access and share files, while corporate IT needs to govern and secure those files. Both parties look hopefully to the cloud to provide the mobility and scalability necessary for this level of file collaboration and control. They are right about cloud mobility and scalability but external cloud services can’t by definition provide IT governed data services and private clouds solutions to-date haven’t exactly been on par with the consumer-grade versions publicly available. End users today demand IT provide services like EFSS that are as good or better then the free services everyone has on their smartphones and laptops, and if IT can’t, end-users will end-run around IT.
The solution isn’t hard to dream up: create a single strategic control point for enterprise-level data mobility and governance that works across enterprise cloud-like services. Sadly, thinking about something does not make it so. Certainly central control is common for specific computing domains. Storage makers create central management consoles for the systems under their control. Virtualization makers create central management for hundreds and thousands of VMs. Backup makers create central control for data replication across remote sites. And the components specific to sharing and protecting file data are common: enterprise file share and sync (EFSS) and cloud gateways. But what has been missing up until now is central control over all those processes in a simple unified manner, especially when supporting a distributed workforce.
CTERA cloud storage gateways have provided file-based protection and mobility since the company stepped onto the cloud-based management scene. Now a major new release further supports and enables integrated end-to-end file and data services. CTERA’s 5.0 version of its Enterprise Data Services Platform powerfully brings together several key capabilities that help IT easily deliver the distributed services their users want and need while ensuring full governance and control. The platform combines their highly available NAS gateway for core and edge services with advanced EFSS and backup. Their seamless integration cements separate solutions into a cohesive CTERA platform. In this report we will examine CTERA 5.0 and its balanced benefits for both user happiness and productivity, and IT governance and control.