Match performance to applications with secure industry standard servers
HPE announced its new Gen10 servers in June 2017, and we’ve eagerly awaited our first shipments because the new features in introduced in this – the tenth iteration of the ProLiant family – represent something of a step change in ease of use, performance, and security.
HPE claims (not unreasonably) that its customers want better ease and speed of deployment, a reduction in latency, and an increase in performance – all three of which create more agility. Customers also desire better pay-as-you-go, flexible consumption models to compete with public cloud, and better security to match today’s rapidly evolving threats. All these considerations have been factored into Gen10.
HPE believes Gen10 servers are key to infrastructure modernisation, and that they’re ideal for accelerating business insights across a hybrid world of traditional IT, public and private cloud. Here we’re going to pick out a few of the stand-out features that caught our eye.
The most secure industry standard server?
HPE’s not backward about coming forward with what it believes is a massive leap in server security. Gen10 servers feature HPE’s unique firmware protections, embedding cybersecurity controls right into the silicon.
Integrated Lights-Out, or iLO, is a proprietary embedded server management technology that enables a “Silicon Root of Trust” to mitigate firmware-based cyber threats. Gen10 servers offer ways to blunt the impact of ransomware, for example, by ensuring you preserve your data and can run your business without the risk of paying criminal hackers.
Andrew Watson, chief technologist for global industries at HPE, said: “One of the weaknesses that we’ve seen in hardware is firmware attacks. We can build in a ‘fingerprint’ into the iLO chip, and unless the fingerprint of that firmware when it boots up matches the fingerprint of the hardware it won’t boot.” HPE claims that as all the technology from the silicon to the firmware is proprietary, the only way a mismatch in this digital fingerprint could happen is if the firmware has been tampered with.
Gen10 also has an Enterprise Secure Key Manager tool built in, which enables organisations to secure cloud, storage, and server together, giving a greater consistency of experience. And Niara (acquired by HPE in February) behavioural monitoring can, in combination with HPE Aruba ClearPass technology, identify and isolate any potentially infected devices that are trying to access the network.
With HPE firmware threat detection, you know whether your firmware has been compromised and in the event of a breach, you can automatically recover to a known trusted state and quickly get your sever operational again. And through implementation of the highest level of security algorithms, known as CNSA-suite, you can keep your most confidential information within your Gen10 server protected.
The HPE infrastructure stack, including servers, storage, and networking, has been validated against the NIST 800-53 controls – foundation for accelerating regulatory compliance initiatives like EU General Data Protection Regulation (EU GDPR).
Alain Andreoli, head of HPE’s infrastructure group, told HPE Discover delegates in June: “It’s all about the life cycle of the data. Security’s a long journey. We have even planned for the grave. When your server is being disposed of, its embedded data cannot be reconstructed or retrieved any longer. We protect it forever.”
Independent analysts Moore Insights and Strategy have produced a white paper discussing HPE’s approach and the Silicon Root of Trust that you can download here.
And you can learn more about the Silicon Root of Trust in HPE’s video.
Scalable Persistent Memory matches compute to workload
Along with Silicon Root of Trust, HPE’s also made quite a song-and-dance about Scalable Persistent Memory, which it says offers “the performance of memory with the persistence of storage”.
Scalable Persistent Memory is an integrated storage solution that runs at memory speeds with terabyte-scale capacity, unlocking new levels of compute performance with built-in persistence.
With up to 27 times faster application checkpoint operations and 20 times faster database restores (based on comparing restoring a Microsoft SQL Hekaton in-memory database with SSDs versus restoring with HPE Scalable Persistent Memory), HPE claims its solution delivers the fastest persistent memory in the market at scale.
HPE’s approach combines standard DRAM along with NAND flash memory and a micro-controller with an integrated battery on a module that fits in a standard memory slot. Unlike the first generation of persistent memory, which was available as 8GB NVDIMMs, the latest generation making its debut with the Gen10 servers is available in 16GB modules. Gen10 servers also will support future memory types including Intel 3X XPoint.
You can watch a short video on HPE Scalable Persistent Memory.
Intelligent System Tuning that avoids the jitters
For several years the base frequency of server CPUs has generally remained stable while the number of cores has increased. A number of processors run opportunistically at higher frequencies when these extra cores or the power to run them are not being utilised.
Although these opportunistic frequency capabilities can increase performance, they also introduce an unwanted side effect. Frequency shifting itself introduces computation ‘jitter’ and undesirable latency. For example, high-frequency traders who rely on time-sensitive transactions can’t tolerate the microseconds of delay that can be added randomly to a trade by a frequency shift, as over time such delays can cost millions of dollars. Consequently the current trend for latency-sensitive customers is to disable features that normally would result in increased application performance.
Starting with Gen10 servers using Intel Xeon Scalable Processors and iLO 5, HPE has introduced a new feature that allows customers to achieve both frequency upside and low latency. Processor Jitter Control allows the customer to remove or reduce the jitter caused by opportunistic frequency management, resulting in better latency response and higher throughput performance.
HPE VP Tim Peters said: “Now, what’s unique about this is, we own the IP here. We work with Intel to unlock registers, to allow us to have that dynamic state and that control.” Intelligent System Tuning can be used to throttle the server as well. “We can dynamically decrease performance as the application workload steps down because not all applications need to be running at the highest performance,” said Peters. “You can step that down and save on energy.”
Read HPE’s white paper on Intelligent System Tuning here.
Rolling out Gen10 server technologies
The HPE Gen10 ProLiant DL family matches a wide range of compute requirements. The 500 series consists of scale-up servers for business critical workloads like Microsoft SQL Server and Oracle environments. The 300 series comprises versatile performance servers to support a variety of applications. The 100 series optimises storage, performance, efficiency, and manageability to address multiple workloads. And the 10 series is right-sized for the budget-conscious.
The latest Gen10 developments in storage and memory technology, will be rolled out to other products too, HPE vice president Tim Peters said, it will be “also on our tower platform. Also on our Synergy platforms … Anything that uses our iLO technology, that’s where you’ll find it because that’s what it’s embedded in.”
HPE’s Gen10 servers also will be a key component of HPE’s Genesis platform for accelerating applications and business insight, providing comprehensive agility, security and simplicity and control, added Peters.
To help with management of the new Gen10 servers, HPE also introduced HPE OneView 3.1 to better support their use in software-defined infrastructures. HPE OneView 3.1 will help Gen10 servers provide composable storage capabilities, and improves firmware management.
Flexible capacity pricing: just pay for what you need
We all know businesses can often end up paying for services they don’t use as part of a business IT purchase. This is one of the arguments often put forward in favour of public cloud.
In response, HPE’s announced that Gen10 servers are available with a pricing option that will allow users to scale up or down while only paying for what they use. Customers can retain an on-site buffer for when demand is high. This means no more overprovisioning and no more upfront cash for something that may not be used for a while.
While this flexible capacity pricing is an ideal solution for larger enterprises, smaller organisations can benefit from HPE’s new Capacity Care service, which takes some of Flexible Capacity’s key components and provides a dashboard through which organisations can monitor when they need to add capacity to their network. This new specification will also ease the procurement process for these firms.
Current users of Gen9 servers will be able to “upgrade to Gen10 with no upgrade in your payments,” said Alain Andreoli, head of HPE’s infrastructure group.
Naturally, Vohkus can advise on the most appropriate financing options for your needs.
If you’re considering an on-premises or hybrid solution, HPE Gen10 servers offer considerable advantages without compromising security, performance, recoverability or even the commercial flexibility demanded in today’s market.
Speak to one of our HPE consultants today to find out how cost-effective an upgrade to Gen10 can be for you.