The healthcare industry continues to face tremendous cost challenges. The U.S. government estimates national health expenditures in the United States accounted for $3.2 trillion last year – nearly 18% of the country’s total GDP. There are many factors that drive up the cost of healthcare, such as the cost of new drug development and hospital readmissions. In addition, there’s compelling studies that show medical organizations will need to evolve their IT environment to curb healthcare costs and improve patient care in new ways, such as cloud-based healthcare models aimed at research community collaboration, coordinated care and remote healthcare delivery.
For example, Goldman Sachs recently predicted that the digital revolution can save $300 billion in spending in the healthcare sector by powering new patient options, such as home-based patient monitoring and patient self-management. Moreover, the most significant progress may come from a medical organization transforming their healthcare data infrastructure. Here’s why:
- Advancements in digital medical imaging has resulted in an explosion of data that sits in picture archiving and communications systems (PACS) and vendor neutral archives (VNAs).
- Patient care initiatives such as personalized medicine and genomics require storing, sharing and analyzing massive amounts of unstructured data.
- Regulations such as the Health Insurance Portability and Accountability Act (HIPAA) require organizations to have policies for long term image retention and business continuity.
Unfortunately, traditional file storage approaches aren’t well-suited to manage vast amounts of unstructured data and present several barriers to modernizing healthcare infrastructure. A recent Taneja Group survey found the top three challenges to be:
- Lack of flexibility: Traditional file storage appliances require dedicated hardware and don’t offer tight integration with collaborative cloud storage environments.
- Poor utilization: Traditional file storage requires too much storage capacity for system fault tolerance, which reduces usable storage.
- Inability to scale: Traditional storage solutions such as RAID-based arrays are gated by controllers and simply aren’t designed to easily expand to petabyte storage levels.
As a result, healthcare organizations are moving to object storage solutions that offer an architecture inherently designed for web scale storage environments. Specifically, object storage offers healthcare organizations the following advantages:
- Simplified management, hardware independence and a choice of deployment options – private, public or hybrid cloud – lowers operational and hardware storage costs
- Web-scale storage platform provides scale as needed and enables a pay as you go model
- Efficient fault tolerance protects against site failures, node failures and multiple disk failures
- Built in security protects against digital and physical breeches
For hospitals and medical research institutes, the ability to interpret genomics data and identify relevant therapies is key to provide better patient care through personalized medicine. Many such organizations are racing forward, analyzing patients’ genomic profiles to match more clinically actionable treatments using artificial intelligence (AI).
These rapid advancements in genomic research and personalized medicine are very exciting, but they are creating enormous data challenges for healthcare and life sciences organizations. High-throughput DNA sequencing machines can now process a human genome in a matter of hours at a cost approaching one thousand dollars. This is a huge drop from a cost of ten million dollars ten years ago and means the decline in genome sequencing cost has outpaced Moore’s Law (see chart). The result is an explosion in genomic data – driving the need for solutions that can affordably and securely store, access, share, analyze and archive enormous amounts of data in a timely manner.
Challenges include moving large volumes of genomic data from cost-effective archival storage to low latency storage for analysis to reduce the time needed to analyze genetic data. Currently, it takes days to do a comprehensive DNA sequence analysis.
Sharing and interpreting vast amounts of unstructured data to find relationships between a patient’s genetic characteristics and potential therapies adds another layer of complexity. Determining connections requires evaluating data across numerous unstructured data sources, such as genomic sequencing data, medical articles, drug information and clinical trial data from multiple sources.
Unfortunately, the traditional file storage within most medical organizations doesn’t meet the needs of modern genomics. These systems can’t accommodate massive amounts of unstructured data and they don’t support both data archival and high-performance compute. They also don’t facilitate broad collaboration. Today, organizations require a new approach to genomics storage, one that enables:
- Scalable and convenient cloud storage to accommodate rapid unstructured data growth
- Seamless integration between affordable unstructured data storage, low latency storage, high performance compute, big data analytics and a cognitive healthcare platform to quickly analyze and find relationships among complex life science data types
- A multi-tenant hybrid cloud to share and collaborate on sensitive patient data and findings
- Privacy and protection to support regulatory compliance
If you are an existing customer of HPE 3PAR, this latest release of 3PAR capabilities will leave you smiling. If you are looking for an All Flash Array (AFA) to transform your data center, now might be the time to take a closer at HPE 3PAR. Since AFAs first emerged on the scene at the turn of this decade, the products have gone through various waves of innovation to achieve the market acceptance it has today. In the first wave, it was all about raw performance for niche applications. In the second wave, it was about making flash more cost effective versus traditional disk-based arrays to broaden economic appeal. Now in the final wave, it is about giving these arrays all the enterprise features and ecosystem support to completely replace all legacy Tier 0/1 arrays still in production today.
HPE 3PAR StoreServ is one of the leading AFAs on the market today. HPE 3PAR uses a modern architectural design that includes multi-controller scalability, a highly-virtualized data layer with three levels of abstraction, system-wide striping, a highly-specialized ASIC and numerous flash innovations. HPE 3PAR engineers pioneered this very efficient architecture well before flash technology became mainstream and proved that this architecture approach has been timeless by demonstrating a seamless transition to support all-flash technology. During this same time, other vendors ran into architectural controller-bound bottlenecks for flash, making them reinvent existing products or completely start from scratch with new architectures.
HPE’s 3PAR timeless architecture has meant that features introduced years before are still relevant today and features introduced today are available to current 3PAR customers that purchased arrays previously. This continuous innovation of features available to old and new customers alike provides the ultimate in investment protection unmatched by most vendors in the industry today. In this Technology Brief, Taneja Group will explore some of the latest developments from HPE that build upon the rich feature set that already exists in the 3PAR architecture. These new features and simplicity enhancements will show that HPE continues to put customer’s investment protection first and continues to expand its capabilities around enterprise-grade business continuity and resilience. The combination of economic value of HPE 3PAR AFAs with years of proven mission critical features promises to accelerate the final wave of the much-anticipated All-Flash Data Center for Tier 0/1 workloads.
We are moving into a new era of data storage. The traditional storage infrastructure that we know (and do not necessarily love) was designed to process and store input from human beings. People input emails, word processing documents and spreadsheets. They created databases and recorded business transactions. Data was stored on tape, workstation hard drives, and over the LAN.
In the second stage of data storage development, humans still produced most content but there was more and more of it, and file sizes got larger and larger. Video and audio, digital imaging, websites streaming entertainment content to millions of users; and no end to data growth. Storage capacity grew to encompass large data volumes and flash became more common in hybrid and all-flash storage systems.
Today, the storage environment has undergone another major change. The major content producers are no longer people, but machines. Storing and processing machine data offers tremendous opportunities: Seismic and weather sensors that may lead to meaningful disaster warnings. Social network diagnostics that display hard evidence of terrorist activity. Connected cars that could slash automotive fatalities. Research breakthroughs around the human brain thanks to advances in microscopy.
However, building storage systems that can store raw machine data and process it is not for the faint of heart. The best solution today is massively scale-out, general purpose NAS. This type of storage system has a single namespace capable of storing billions of differently sized files, linearly scales performance and capacity, and offers data-awareness and real-time analytics using extended metadata.
There are a very few vendors in the world today who offer this solution. One of them is Qumulo. Qumulo’s mission is to provide high volume storage to business and scientific environments that produce massive volumes of machine data.
To gauge how well Qumulo works in the real world of big data, we spoke with six customers from life sciences, media and entertainment, telco/cable/satellite, higher education and the automotive industries. Each customer deals with massive machine-generated data and uses Qumulo to store, manage, and curate mission-critical data volumes 24x7. Customers cited five major benefits to Qumulo: massive scalability, high performance, data-awareness and analytics, extreme reliability, and top-flight customer support.
Read on to see how Qumulo supports large-scale data storage and processing in these mission-critical, intensive machine data environments.
Every year Dell measures the availability level of its Storage Center Series of products by analyzing the actual failure data in the field. For the past few years Dell has asked Taneja Group to audit the results to ensure that these systems were indeed meeting the celebrated 5 9s availability levels. And they have. This year Dell asked us to audit the results specifically on the relatively new model, SC4020.
Even though the SC4020 is a lower cost member of the SC family, it meets 5 9s criteria just like its bigger family members. Dell did not cut costs by sacrificing availability, but by space-saving design like a single enclosure for media and controllers instead of two separate enclosures. Even with the smaller footprint – 2U to the SC8000’s 6U -- the SC4020 still achieves 5 9s using the same strict test measurement criteria.
Frankly, many vendors choose not to subject their lower cost models to 5 9s testing. The vendor may not have put a lot of development dollars into the lower cost product in an effort to reduce cost and maintain profitability on a lower-priced system.
Dell didn’t do it this way with the SC4020. Instead of watering it down by stripping features, they architected high efficiency into a smaller footprint. The resulting array is smaller and more affordable, and retains the SC Series enterprise features: high availability and reliability, performance, centralized management, not only across all SC models but also across the Dell EqualLogic PS and FS models. This level of availability and efficiency makes the SC4020 an economical and highly efficient system for the mid-market and the distributed enterprise.
Deduplication is a foundational technology for efficient backup and recovery. Vendors may argue over product features – where to dedupe, how much capacity savings, how fast are its backup speeds -- but everyone knows how central dedupe is to backup success.
However, serious pressures are forcing changes to the backup infrastructure and dedupe technologies. Explosive data growth is changing the whole market landscape as IT struggles with bloated backup windows, higher storage expenses, and increased management overhead. These pain points are driving real progress: replacing backup silos with expanded data protection platforms. These comprehensive systems backup from multiple sources to distributed storage targets, with single console management for increased control.
Dedupe is a critical factor in this scenario, but not in its conventional form as a point solution. Traditional dedupe is suited to backup silos. Moving deduped data outside the system requires rehydrating, which impacts performance and capacity between the data center, ROBO, DR sites and the cloud. Dedupe must expand its feature set in order to serve next generation backup platforms.
A few vendors have introduced new dedupe technologies but most of them are still tied to specific physical backup storage systems and appliances. Of course there is nothing wrong with leveraging hardware and software to increase sales, but storage system-specific dedupe means that data must rehydrate whenever it moves beyond the system. This leaves the business with all the performance and capacity disadvantages the infrastructure had before.
Federating dedupe across systems goes a long way to solve that problem. HPE StoreOnce extends consistent dedupe across the infrastructure. Only HPE provides customers deployment flexibility to implement the same deduplication technology in four places: target appliance, backup/media server, application source and virtual machine. This enables data to move freely between physical and virtual platforms and source and target machines without the need to rehydrate.
This paper will describe the challenges of data protection in the face of huge data growth, why dedupe is critical to meeting the challenges and how HPE is achieving the vision of federated dedupe with StoreOnce.