Interview: Robin Hughes, Secure Logiq
One of my customers approached me and explained that by optimising the server hardware specifically for the HD CCTV application it was being used for you can massively increase the performance. He proved to me that he could achieve over 15 times the incoming camera throughput of the servers and we decided that there was a business opportunity. It is Secure Logiq’s tenth anniversary this year and I am delighted to say we were right.
a&s Adria: Mr. Hughes, for a start, could you present yourself, your educational and professional background, and the role you’re playing at Secure Logiq?
Hughes: I’m Robin Hughes, co-founder of Secure Logiq and known to the industry as an evangelist for the technology we make. As a company that has grown organically over the last ten years my various roles have included pre-sales, sales, marketing, system design and even accounts, in fact anything that doesn’t involve using a screwdriver.
I have worked in the electronic security product industry since I left university nearly 30 years ago, the vast majority of this time in the distribution and manufacturing sector. I have been lucky enough to have been at the cutting edge of CCTV innovation for my whole career, being the first to introduce colour/mono cameras, speed domes, digital video recorders, Video analytics and Megapixel CCTV to the EMEA market. I have been at the forefront of, and consider myself to have played a big part in the transition of the CCTV industry from analogue to digital to megapixel where the requirement for ultra-high throughput, high storage density servers began. Whilst I am by no means a technical person, I have always made a point of knowing a product inside out before I promote it and have an excellent understanding of integration and system design. All of these traits I believe are necessary to get a full understanding of the amazing industry we are part of.
a&s Adria: Since 2011, Secure Logiq has been developing a missing piece of the HD surveillance jigsaw. What is that piece and how did you and your colleagues come up with the idea to fill that gap?
Hughes: The transition from analogue to IP surveillance has been incredibly slow, mainly due to the difference in skill set between the physical security integrator and IT infrastructure specialists. For many reasons our industry has always been a bit behind the curve when it comes to IT expertise. I was working for well-known megapixel camera and VMS manufacturer who were using commercial off the shelf (COTS) servers as the recording medium. One of my customers who was purchasing cameras and VMS software only approached me and explained that by optimising the server hardware specifically for the application it was being used for (in this case HD CCTV) you can massively increase the performance. He proved to me that he could achieve over 15 times the incoming camera throughput of the servers I was selling at the time as an end-to-end solution and we decided between us that there was a massive gap in the market and a business opportunity. It is Secure Logiq’s tenth anniversary this year and I am delighted to say we were right and are now the leading independent manufacturer of servers optimised for HD surveillance globally.
a&s Adria: Users can select best-of-breed IP cameras or VMS, but then select a server that was never designed and optimized for video. What are the shortcomings of that?
Hughes: We have server, workstation and client hardware for every application. In all honesty, for most small to medium sized installations which are not utilising a DVR / NVR solution a COTS server will handle the data just fine. However, when you get larger or more complex projects that is when that deep knowledge of everything involved with HD surveillance kicks in.
The mathematics behind calculating the processing and storage requirements for video is actually quite simple. It is based on a few basic variables such as camera resolution, frame rate and the archive time required. Where it gets more complicated is when you have to start factoring in additional functionality such as licence plate recognition, advanced video analytics, streaming to mobile devices, or thinking about the outgoing bandwidth requirements for a large control room with multiple monitors. Add to this that every piece of software operates in a different way and interacts and utilises different components and you soon discover that specialist knowledge is required. Within the IT world, software and hardware are also completely different disciplines which is why to date most VMS suppliers are content with recommending or rebadging IT centric solutions rather than a specialist optimised product. However, a well-designed hardware solution can offer savings in capital expenditure whilst significantly reducing annual operational costs.
a&s Adria: When we talk about performance, what’s the top speed your servers can achieve? What would you normally use to test it and what tool do you use instead?
Hughes: There is a common misconception in the industry that to increase the total incoming camera throughput of a server you simply need to upgrade the processor. However, with specialist knowledge and the correct selection of components there are much better ways to optimise throughput. In our industry there is a huge and often confusing variation in the way that the processing capability is defined. It could be number of cameras, total number of megapixels, average bandwidth or many more, but this still leaves so many variables that mistakes are often made and the server will underperform or even fail. The only way to accurately represent this number is the total peak camera traffic that the server can handle with certain software and whilst performing multiple tasks but without having to make compromises on camera performance. Our highest performing servers are available today with incoming camera throughputs in excess of 4,000Mbps, equivalent to more than ten IT-centric servers. When we released these numbers to the industry, obviously we were asked to ‘prove it’ which at the time would have involved setting up 1,000 HD cameras with a rack full of switches which was unrealistic so our software development department engineered a solution called Logiqal Benchmark, a software which allows us to create a virtual IP CCTV system of any size.
a&s Adria: Do you use Logiqal Benchmark simulator to optimize the system for different VMS solutions? Can you give us an example?
Hughes: Logiqal Benchmark contains a ‘camera replicator facility’ allowing us to take any number of live or pre-recorded streams of video and duplicate them multiple times on distinct IP addresses to simulate the load of all cameras recording to an IP surveillance system of any size. Additional system features such as multiple client machines, ANPR, video analytics and mobile transcoding can also be added virtually to give a simulation of the overall system and our advanced diagnostics will provide valuable and accurate system data, such as actual and peak throughput ratings, disk read / write speeds, power consumption and temperature readings.
Different VMS solutions can be applied to compare the various software’s efficiency with the specified hardware solution. We use this as a tool to monitor how the data interacts with the server and fine tune components and settings to optimise the server towards a particular VMS. The server we sell for an Avigilon solution might be very different to a server we sell for a Milestone application even with the exact same camera count, resolution and frame rates. Logiqal Benchmark has proven to be a great tool for our internal testing but also for allowing us to prove that we can do a lot more with a lot less hardware therefore minimising cost, power and rackspace requirements.
a&s Adria: Since the law in countries around the world regulate the length of archiving recordings differently, what can Secure Logiq do to offer more storage space?
Hughes: We do have many more options available than traditional IT distributors to offer larger storage on a scale that is not often demanded in typical IT scenarios. There is nothing more storage intensive than real-time video other than multiple streams of HD video recording simultaneously for extended periods of time. As a result, we need much more storage than typical server room applications. At Secure Logiq, we offer the industry’s highest storage density servers and currently are able to offer direct attached storage (DAS) options of over 1.5 Petabytes in a single 4U server. This is of course a large number but remember that different regions demand different storage archives.
Due to data protection laws, large parts of Europe only record video for seven days in a typical installation. In the UK and most of the world 30 days is commonly specified, but in the Middle East it is not unusual to record the video for 90 or even 180 days which would be exactly three or six times the storage required for the same project in the UK. For even larger storage requirements we also offer SAN solutions that can scale to over 12 Petabytes on a single recording server. Most of our customers rely on us to help them determine the best storage solution for their application.
a&s Adria: It’s interesting that you have tested how much storage you need with different industry storage calculators and the same propositions. What results did you receive?
Hughes: In general, industry calculators are poor and often give wildly varied results and partial information. As an experiment I inputted the exact same camera statistics into five different well-known industry calculators, it was 100 x 1080P cameras recording with H264 compression, default quality, medium scene complexity at 12 frames per second for 31 days. Results varied from 49TB to 175TB with the lower numbers interestingly being from VMS brands that did not supply their own server hardware.
Technically the mathematics is exactly the same with the only variables being an individual’s concept of what constitutes an average acceptable image and how they have defined how complex that image is at the programming level. This is where vast experience comes into play. At Secure Logiq we have invested heavily in our design team and offer a design guarantee, if you let us do the calculations for you, we will guarantee the processing and storage necessary and if we get it wrong we make up any shortfall ourselves, free of charge.
a&s Adria: What RAID architecture do you use to achieve greater resilience against data loss?
Hughes: To help redundancy and improve performance the Secure Logiq range of HD servers utilise a number of different RAID architecture levels. The type of RAID structure we use depends on what is best for your application. We can supply any RAID configuration, but our default configuration is optimised for value and performance versus risk. In general terms, the more hard drives we utilise in a system the more redundant drives we have. Our best-selling servers which are the 2U servers do utilise RAID5 + Hot spare which like RAID6 has two ‘redundant’ HDDs but in the event of a disk failure has a significantly faster rebuild time, which is very important when your servers are under as much stress as an HD CCTV application. When correctly configured this will also increase the total throughput of the unit. Our ultra-high density SAN units offer erasure coding which has advantages over RAID in terms of rebuild times and flexibility when expanding the storage of the system whilst retaining the existing archive.
a&s Adria: What are the other features of Secure Logiq servers that increase their resilience? Why do you use enterprise drives instead of surveillance drives?
Hughes: If you are using a VMS and a server rather than an NVR solution, then it is likely that you are in the enterprise CCTV arena, and it is also likely that the solution is for a mission critical application. Based on this assumption, all Secure Logiq servers are built with multiple layers of resilience as standard, so that they do not fail in the field or if a component fails it will not affect the overall system operation.
All of our enterprise server range is hand made in London with a strict quality control process during which all products are soak and load tested before despatch. We only utilise the highest quality enterprise components and all servers feature a five-year, next business day, global onsite warranty. Often offered as optional extras by our competitors, we design in resilience features as standard, such as dual redundant power supplies; the Operating System and VMS applications separated from the main HDD archive on two mirrored Solid State Drives (SSD’s) in RAID1; and Battery backed RAID to ensure the disk array shuts down correctly in case of power failures. As mentioned before, we offer RAID to suit the application, IPMI for advanced remote diagnostics and support and Logiqal Healthcheck for system status monitoring.
Regarding the drives, if you go to any HDD manufacturer and ask what you should use to store mission critical data, they will point you to Enterprise HDDs every time. Our industry is ‘Big Data’, nothing puts HDDs under as much duress as Gigabits of data constantly writing and overwriting, so you want something that was designed for this purpose. Enterprise drives are rated for continuous use with a five-year warranty, they are designed for high throughput (540TB per annum), they have advanced enterprise class architecture which leads to a low failure rate which is ultimately best for our customers but also reduces our support burden.
a&s Adria: You are the only CCTV partner for Stratus everRun in the EMEA region. What are the benefits of having two mirrored management servers?
Hughes: In the VMS world there are two types of system, those with intelligence and functionality distributed between the server estate (Avigilon and Axxonsoft for example) and those with a management server (Milestone and Genetec). Whilst both have their benefits, and all offer advanced failover models, in the second model the management server becomes the single point of failure with the potential to take an entire system offline.
Having two mirrored management servers utilising Stratus everRun eliminates that risk without the need for very advanced IT skills and additional Microsoft licencing. People often forget that even with failover on the recording servers, if you lose a server you lose the video archive on that server so you will not have access to the video from the cameras stored on those servers. However, in tandem with Stratus everRun we can also help design advanced failover models like ‘dual recording’ so that even when losing a management server and a recording server we would have zero downtime and the customer can continue to access the video archive from all of the cameras for the most recent events. We have deployed these systems in multiple high security applications such as banks, airports and datacentres.
a&s Adria: All of your servers come with Logiqal Healthcheck Pro. What processes do you monitor with that tool and can it help to prevent storage failures?
Hughes: Written internally by our software development team, Logiqal Healthcheck Pro is a software utility which is installed on all Secure Logiq hardware and, given permission, will monitor every process and component within our servers. This allows us and our customers to reduce maintenance requirements whilst increasing support performance. Our customers can view their entire server estate globally, either displayed as icons or on a map, and a full set of green lights will give them the confidence that there are no issues on site. A simple to follow traffic light system will alert customers of any potential or actual problems within the server, down to the component level including failed hard drives or temperature issues. The system can also be configured to send emails or SMS alerts depending on the severity of the problem so that an engineer can be alerted within five seconds of any issue arising. This allows our customers to offer a level of pre-emptive maintenance and support which has never been seen before in the industry
a&s Adria: Maybe you could share with us a case study so we can grasp all the benefits of customized Secure Logiq servers for video surveillance systems?
Hughes: Let me give you an example of surveillance solution that benefited purely from upgrading the server element of the solution. A well-known London-based English Premier League football club were early adopters of HD surveillance and immediately understood the benefits of an ultra-high resolution surveillance system which featured large sensor multi-megapixel cameras installed across the entire stadium. Six years ago, storage was a premium and three quotations were offered by the integrator £150,000, £300,000 and £1M, the only difference between these quotes was the total amount of processing and storage that was on offer. The end user opted for the middle quotation.
The IT world moves much faster than the CCTV industry and it is commonly accepted that IT hardware is ‘end of life’ after five years while modern solid state cameras have a lifespan of 10-15 years and VMS software usually has a paid or free upgrade path for the life of the system, though this again is highly dependent on the suitability of hardware being used. When the existing servers began to fail, and the software demanded more modern technology to fully utilise the latest features the integrator approached Secure Logiq for advice. We were able to replace the existing 15 servers with just two, high throughput, high storage density servers but at the same time offer them six times the processing and storage than they previously had.
The end result was that the football club was able to increase the frame rate to its maximum on every existing camera at the stadium and massively reduce compression, therefore increasing image quality. This completely transformed the existing security system by simply replacing the server hardware but in the time elapsed from the original installation the price for the server hardware was a fraction of the top level quotation.
a&s Adria: You recently attended Adria Security Summit in Belgrade as an exhibitor. Are you looking for regional partners and what are your plans for the Southeast Europe security market in general?
Hughes: We are constantly striving to increase our global footprint and our debut at the Adria Security summit this year was a really great solution for increasing our visibility in this region. We currently export to and support products in 42 countries across six continents and we are hoping to attract both integration and distribution partners in Southeast Europe and beyond.
This is one of the main reasons why we hired an extremely experienced and well-known European Sales Manager, Ivan Sval, who is based in Slovenia. We are certain that not just through our product but our extensive knowledge of IP CCTV and our incredible design services, we will be able to add value and help our partners succeed with any project utilising a VMS or video analytics-based solution.
a&s Adria: What can we expect from Secure Logiq in the years ahead, in terms of expansion of the product portfolio, new business strategies, and technologies?
Hughes: Video Analytics is an area we are heavily invested in, which is a really specialist area of server design. As these features become more cost effective and even free in some enterprise VMS solutions the server design will have increased importance even in the small to medium size projects. These applications will nearly always require a custom designed server and returning the designs quickly as well as being able to build and ship the equipment in a short time frame will remain our prime objective in supporting our partners.
As you have heard we have already invested heavily in a software development team, but our product management department are really excited to create hardware elements specifically designed for the efficient processing of HD video. Our primary growth objective is to replicate the success we have seen in the UK and many other regions by attracting a larger global customer base to join the Secure Logiq family.