Isn't it about time someone brought true innovation back to workflow storage?
We think it is.
The virtualized RAID kernel is state of the art and delivers not only fast consistent data delivery, but offers a host of
optimizations that provide better performance in media-centric operations. The ability of the RAID kernel to work with
the hardware to provide real-Active -Active load balancing and automated fail over and fail back allows Symply to
deliver the best performance in non-perfect-state instances. Additionally the plug-in architecture provides TRIM and
performance and reliability profiles for mixed SSD/HDD configurations and maintains performance and reliability
standards across both types of media. When combined with Enterprise-class Hard drives and SSD’s (the only type we
use) the performance of the system stays consistent day to day, month to month and year to year, accounting for
aging components and differing conditions.
The greatest differentiators unique to Symply come from the addition of patented licensed technologies
that greatly enhance the performance, high-availability and the overall the capabilities of the Symply platform.
Fast Rebuild - The Symply platform is unmatched in drive rebuild speed and performance levels in degraded and
rebuild states. While others nominally rebuild large drives too slow to make large fast RAID sets practicable,
Symply makes it easy. Compare typical rebuild times from other vendors serving the Media markets. Typically they
will rebuild an 8TB hard drive in 96-120 hours depending on load. Symply will commonly rebuild at 1TB per hour
under normal load. or 8TB in 8 hours. Well within the nightly window when most production facilities can provide
lighter loads. Under no load its even faster.
Advanced Protection - The use of RAID 6 is common in most RAID-based enterprise class systems today in order to
provide protection against data loss in the event of a failed drive. The reason RAID 6 (double parity) is used is
because the old standard RAID 5 was deemed “too dangerous” with larger hard drives and their extremely long
rebuild times. These long rebuild times left open the possibility of a “cascading drive failure”, a relatively common
possibility when an older hard drive fails and rest of the RAID set is asked to take over its load and also the
increased load of rebuild. In a lot of cases the drives are driven beyond 100% duty (for the entire and long rebuild
cycle) and this causes a second drive to fail. In the case of RAID 5 with a cascading drive failure, all your data
(content) would be lost. This made the case for RAID 6. to become the standard. In other RAID systems (non
Symply) by using RAID 6 all you have done is bought more time (maybe) as during every drive rebuild, you still have
placed undue load on the rest of the drives in your system during a rebuild and have actually increased the
possibility for a cascading failure in the future. With Symply, the fast rebuild feature decreases the load on the RAID
set and greatly reduces the possibility of a cascading failure during rebuild and in the future. But more importantly,
it allows you to get back to full speed operation faster eliminating any workflow slow downs.
When Advanced Protection is not enough, go “AdvancedPlus” - Advanced Protection offers significant advantages
over other offerings, but what if your content is not archived yet, or backed up and you want better protection.
Symply has an answer. Borrowing from the advantages of object storage without the performance penalties,
Symply’s AdvancedPlus Protection takes advantage of the unique nature of media deployment structures. Content
creators use lots of space and its only getting bigger. With 4K and beyond and digital cameras always capturing, 1,
2, 5 and even 10TB shoots are commonplace. The size of typical WIP (work in process) workspaces can exceed
500TB and that takes a lot of hard drives to accomplish that capacity. We have been able to leverage the number of
RAID sets (sets of hard drives) to deliver a significant protection advantage that does not compromise
performance. This patented technology allows us to carve a small portion from each RAID set and dedicate it as a
“parity stripe” for all the remaining RAID sets in the workspace providing virtualized protection against multiple hard
drive failures in multiple RAID sets with no data loss, and in most cases, no performance hit. While overall capacity
will be somewhat affected, this extra protection provides assurance that your valuable content is always safe, and
does not require a second costly near-line archive to supplement to offline and/or offsite archive. This single feature
make Symply the most resilient integrated storage workspace available. Nothing in our price range or even far
above, can offer comparable failed-drive and rebuild performance or protection. We care about protecting content
and relieving the stress of any performance issues when impending deadlines approach.
The Symply Advantage is easy to demonstrate. As a company Symply combines years of leading storage innovation and design, with an intimate knowledge of media workflow. We have added world-class support and award-winning software design to round out a workflow storage expert team.
The team at Symply is hand-picked. Some members of the team have worked together for years, always striving to further the evolution of workflow, while others have joined this venture to add fresh ideas and new knowledge to our DNA. All in all, we live to innovate and deliver the best user experience and most reliable workflow storage possible
while delivering peak performance to meet the ever changing needs of media professionals.
Our Simple Philosophy – We are not locked to the past, but we look to improve the future
Hardware – Understanding our strengths and weaknesses from the beginning, and with years of hardware development under our belt, we knew developing our own hardware for Symply was both expensive and a lime consuming process. We knew we had to decide between building a custom hardware solution from the ground up, or partner with a manufacturer to deliver our best-of-breed hardware. At first we were skeptical that any manufacturer could meet our high standards, in fact many we started to work with did not.
We made it a mission to evaluate hardware developers and products. Our reputation brought us an exclusive opportunity by a world-class hardware manufacturer to work with us to develop just what we were looking for, a hardware platform that is not only fast, elegant, and affordable, but what we feel is the only 99.999% certified uptime storage application platform anywhere near our retail price.
This hardware platform and its associated components make up the SymplyULTRA and SYMPLYIO product lines and deliver unmatched versatility and features.
Solution Challenge – Starting with a world-class hardware platform manufactured for us by one of the top three hardware manufactures in the industry, our challenge was to take this platform and make it the best workflow storage solution for the next generation of media production.
This presented a number of challenges: The first being that the market is changing, and fast. Third generation digital production can be maddening. Tighter deadlines, a greater number of ingest points and types, a greater number of intra-WIP operations - transcodes, archives, MAM operations, VFX, VR360, color correction, audio sweetening, HFR, HDR and of course ProRes RAW, 4K, 6K, 8K RAW and in several CODECs - and mostly the need to do all of this at-once in a highly collaboration fashion. Add to this that standard IT technology is rapidly coming into media’s “once private sandbox” and you have some real challenges. Many storage providers fall into two camps on this:
They embrace it fully and move standard NAS into the workflow, or
They continue to deliver “brute forcing” older technology by interspersing developments such as SSD and one-off file systems.
But in the end neither of these approaches moves the evolution forward nor satisfies the user community.
Embracing common enterprise IT file-sharing technology has advantages in reducing cost, complexity, and the requirement for specialized IT staff , but generally it cannot be adopted for all media operations, especially for intense WIP operations.
Many times facilities are swayed by the seemingly simple appeal of the “pure NAS” approach that ends up being a disappointment after it is deployed. The lack of real-time collaboration and uncontrolled latency creates a “workflow back-step” that is often detrimental to consistently meeting the needs of content creation in the new post world defined by high-resolution, high turnover projects.
The back-step comes in the form of placing IP standardization and cost structure over collaboration capability and general workflow best practices.
The key to effective collaboration using media creation tools, remains having enough dedicated bandwidth for delivering the stream count needed combined with low enough latency to eliminate drop frames and performance glitches. In addition, the system must be resilient enough to handle “non-perfect state” operation, such as drive fails, controller fails without affecting normal operations. Beyond this, the system design much account for long periods of 100% duty cycle operation without interruption. This is commonly addressed by increasing the headroom available for crunch-time operations.
Unfortunately both hardware and software designs that are perfectly suited for financial transactions, OLTP, database, cloud, and general business, are not well suited for media production. Because the software designs are focused on the file sizes that generally are far smaller, and the test cases involve high transaction rates (IOPs) rather than uninterrupted streams, hardware designs optimize for this as well, and the resulting general purpose ethernet-based IT storage solution delivers inconsistent results in media production.
This usually is exhibited by intermittent problems when multiple users are performing concurrent different operations such as ingest and editing. These slow downs are compounded when any write operations such as transcode and archive are added.
In order to deal with this, many workgroups “Back-Step” their workflows by adding local fast storage to workstations, or add SSD-based caches to the main workflow storage. While both these approaches can help bandage the problem, they add a host of issues media production eliminated several years ago - version control, multiple copies, copy time, and non centralized storage. In the end this back-step can save money but may also add severe consequences overtime.
We took a different approach
Performance and Capacity Scalability –Traditional SAN solutions have done a good job in providing capacity scalability, and depending on the collaboration software used, some provide good performance scalability. However until now none have provided the scale-out capability that is common with highest-end NAS solutions. With all the disadvantages of pure NAS for media, the scale-out nature of the highest end solutions is still appealing. Symply now brings the best of real scale-out to the Media SAN. Others will use the term scale-out SAN but their definition is very basic. It consists of the additive effect of multiple “individual” hosted LUNs (RAID sets) using the Fibre Channel fabric to connect them together. Each LUN set or RAID controller or pair work independently across the bandwidth sensitive client fabric, and the “sharing” software logically stitches them together. For years this is how we all did it and this is how most do it today. Some years ago, the same team that is at Symply pioneered a “back channel” fabric approach to link individual systems together and stitch the controllers via hardware rather than software, thus eliminating the load on the client facing fabric and providing performance scalability with out sacrificing client latency. Today Symply takes that approach a step further. By utilizing a high-speed back channel built into the Symply hardware, we can link several SymplyIO systems to a single SymplyULTRA and scale to over 35GB/s bandwidth while maintaining the same low latency, making no compromise scale-out a reality for high resolution and high frame rate editing, VFX, and finishing. The link technology is fully integrated into the virtualized RAID kernel and operation is completely native. Because this approach allows the controllers to “bond”, the loads are shared and multi-level active active operation is achieved. This unique technology allows linear and non-linear performance/capacity scaling to meet the individual needs of production facilities and offers maximum versatility when designing and upgrading facilities to handle varying and changing work loads. Best of all, because the systems link together, SymplyRTM - Real Time Management software provides real time analytics of the health and performance as individual systems and as a complete cluster.
Management, Optimization, and Support – The team at Symply has had many firsts in the industry over their career and have been responsible for, or contributed to dozens of patents associated with storage technology and workflow management. At Symply this has not changed. Continuing on the mission we started several years ago when we pioneered statistical analytics for workflow storage, our use of virtualized design has allowed us to take this the next logical step.
Management – From the management perspective we have developed SymplyRTM, a simple to use single-pane-of-glass
at-a-view tool for keeping track of your entire workspace. Deployment and daily monitoring chores are easy. RTM is the user facing component of the software most users only need to view. It’s what is behind SymplyRTM that is really unique and highly valuable to anyone maintaining an active production environment.
Support – At Symply we were not at all happy with the level of support offered in our industry, at least as it exists elsewhere. While once you get to the right person or paid a significant amount yearly to “concierge” support it was acceptable, For a majority of facilities, the tight budgets, production infrastructure and deadline dependency do not allow for dedicated IT workflow storage specialists on site 24/7 to manage issues, nor is there time in the production schedule to perform extensive tests to determine problem root causes. In every case the best outcome is simply replacing a hardware component or installing a software fix, and doing it while the system keeps running. The best support organizations are only as good as the tools they have to work with. Whether at the user site or at their headquarters. To this end we wanted to ensure that our users and technicians had the tools they needed to get the job done fast, efficiently, and with repeatability. This led us to develop a multi-pronged approach which we feel addresses the most hated aspects of supporting workflow under the pressures of production deadlines.
Our approach again leaned heavily on our virtualized design. The unique aspect of a true virtualized design using a Hypervisor and number of PCIe connected Virtual machines is that we run logs and reports on all virtual machines and gather them with another virtual machine and send them where we need to, as-often-as-needed and without an impact on the rest of the system’s operations. So thats what we did. We gather logs from the operating system(s) the Hypervisor, the Applications including StorNext, the RAID Kernel, the hard drives and the IO ports, and we package this up, encrypt it and send it to the Nebula, the Symply Nebula that is - the one that lives in Cyberspace. In actuality, the Symply Nebula lives in AWS (Amazon Web Services) in protected space where every 60 seconds (this can vary) the logs are captured, cataloged and analyzed for your deployment. This can be a single SymplyULTRA, or an entire Workspace with SymplyULTRA, SymplyIO and SymplyULTRA Expansions. By using machine learning and AI, we can determine the health of the systems and measure against the perfect state, and in a next-generation manner, only tell you whats wrong, if something is wrong. We won’t deliver you a bunch of graphs and charts you need to go figure out. This also allows us to tailor the requirements based on your workflow, and keep it 100% media-centric. This tells us (and you) the hardware, the software, and the configuration is performing optimally or maybe we need to tune it up. In addition you get to see what the headroom is and how much harder you can push the workspace when needed. This approach is really the next step from our pioneering and often copied “analytics” graphs and charts. For us the next step is self-healing and optimizing. This is possible by again leveraging the virtual design of the system.
Let’s look at a support example and see how this approach and some other unique Symply logic
has really driven support into the 21st century.
In this example while highly unlikely, let us assume a system power supply has failed. In most systems today you would as an admin likely get a beep on the system and some type of message. Commonly that is an email with a cryptic message. You then search for the support phone number, call the company, transfer to support, and then give that person that your name, company, model, serial number,. You continue and explain the error message,….and then you likely get transferred to the next level tech for trouble shooting, causing you to repeat all that information over again with the new person. If you are lucky you will get and RMA and a new part will be on its way once you give them the address, the phone number and sometimes a credit card number or PO. We have all done this. And because if it, we sometimes buy the higher level of support or even “phone home” but that can be very costly and mostly it may just eliminate a step or two.
This is “acceptable” to most people and feels pretty standard, and as long as your system is 100% redundant you can get away with this support, however, if you use it too much or have complex issue, that runs you through several levels and/or several days, then it can become unbearable.
Personally, we hate this type of support - it’s overly manual, overly complex, and just out of place today. We thought about a better way to it, and here is what we came with…
The Symply approach takes advantage of the powerful redundant computer you have in the Symply platform. We already know who you are (you are logged into SymplyRTM), we have all the logs, we know the location, the model, the serial number and minute-by-minute health of your system and like a secure iOT (internet of things) when we detect a failure we can act, and act quickly. Now let's replay the power supply failure on a SymplyULTRA.
The first thing that is likely to happen is that you will get two simultaneous emails or texts (your choice). The first will be our server alerting you that a power supply has failed and assigning you a “service reference number”. The second will alert you a replacement part is on its and way and confirm the shipping address and information. In some cases depending on your location it will have information for the UBER driver who is dropping the package off. This service is being built out and currently is only available in Los Angeles although we are adding additional cities quickly. If you do nothing, a replacement part will show up with replacement instructions. Since the system is 100% redundant - it’s a hot swap and SymplyRTM will bring it online and report its health to us. If for some reason you saw the first email and quickly called us on the number in the email - we would ask for your reference number, confirm back your name and company, and tell you the part is on the way and give you tracking info. We think this is how support should be done.
Cloud Integration, Remote Locations, Multi-stage WIP – Modern workflows are not limited to a single workgroup or a
single location. Content is acquired remotely and delivered remotely. Work in process workspaces are not all SAN or
NAS or even all high speed. The best technology is used to balance cost, performance, and security.
The symply platform was designed natively to accommodate all of the above. We designed it to deliver outstanding
performance and deliver the capabilities that meet the needs of modern workflow. Let’s review some of the unique
ways Symply addresses these needs.
1. Multiple workgroups or workgroups in multiple locations – Symply Conveyor is an integrated high-speed, accelerated
sync and transfer tool that all hosted VM’s and third party applications can use to keep systems in Sync or transfer
jobs securely building to building, state to state or country to country.
2. Remote ingest and content delivery – Symply Conveyor is not locked to Symply to Symply Sync and transfer. Any
number of end points can be configured easily.
3. Multi-stage WIP and the cloud – more and more facilities are looking to lower costs using smaller on-line WIP
storage and larger low-cost near-line storage. Symply Conveyor using the integrated back channel communication
can transfer from Symply on-line WIP to Symply near-line archive very quickly with no impact on the client fabric.
And as required, can also sync and transfer to the cloud (Amazon S3 now, others later). This functionality is being
added to our API and will be available to eco-system partners as well, MAM, PAM and Archive software can take
direct advantage of these integrated services.
The Best-Of-Breed Approach – It is very true that there are many ways to get to the same end point. In our case that
means assuring on-going compatibility with commonly used media acquisition, production and delivery tools,
delivering top performance, scalability, reliability, and high-availability optimized for current and future workflows.
That is a high bar, and one that we believe no single company can achieve today when you consider the changing and
expanding workflow aspects of our market. Any company that maintains a “not invented here” approach, or utilizes
inferior technology that is not used in the forefront of the industry will never be a leader.
That is why at Symply we have partnered with the best-of-breed partners in the industry and constantly test to
maintain compatibility with the latest companion products used in modern workflows.
While initially it may be less expensive and easier to use second-rate components, users will quickly see the flaws not
only in those products, but also in your logic. This is why we chose to develop Symply products using the best tools in
Hardware – Top five storage manufacturer, top three electronic manufacturer in the world, certified 99.999% uptime
Collaboration – StorNext, undisputed SAN file system for media - 100% Xsan compatible, used by top Studios
Ecosystem – Full access to the StorNext ecosystem, recommended by over 50 top media product vendors
Software – Linux, Rock solid CentOS 7 - KVM Hypervisor, Enterprise class VM - AWS EC2 and S3, for Nebula