Join Transform 2021 for a very powerful themes in enterprise AI & Data. Learn extra.


This article is a part of the Technology Insight sequence, made doable with funding from Intel.


For the final two years, we’ve been writing about Intel Optane persistent memory, figuring out that scores of evaluations and pilot tasks are underway in enterprises world wide. But quotable case research have been uncommon. That modified lately when Intel publicly provided examples of Optane’s real-world suitability and worth.

During a high-profile occasion, Boeing revealed its manufacturing information heart deployment of Intel Optane persistent memory by way of Oracle’s Exadata X8M database servers. Not lengthy after, VentureBeat snagged an unique assembly with Maruti Sharma, Boeing’s chief architect for digital frequent companies and an affiliate technical fellow throughout the aviation big. Sharma manages Boeing’s enterprise databases and different information administration companies. He’s the perfect individual to take us inside this venture and reveal the hands-on particulars of an enterprise-scale Optane deployment.

Key Points

  •  Boeing’s hundreds of Oracle databases have been foundering on outdated commodity servers, whereas information calls for stored growing.
  • The firm determined to consolidate practically 100 database and storage servers onto solely three racks of Oracle Exadata X8M methods supporting Intel Optane persistent memory.
  • The transfer helped drive efficiency enhancements of 2x to 10x in Boeing’s database operations.

    Above: Maruti Sharma, Boeing’s chief architect for digital frequent companies

Boeing’s want

Boeing manufactures on each U.S. coasts, which helps drive the corporate’s want for cloud-based information administration. All informed, Boeing runs over 21,000 databases, roughly 5,000 of that are Oracle. Given how tempting it’s to deploy databases with devoted server stacks, it’s simple to think about sprawl and redundancy accumulating over time. So, as in lots of enterprises, lowering the information heart server footprint is one among Sharma’s high targets.

Not surprisingly, given the large variety of databases, Boeing had an ocean of data it may barely sip. An instance:

“Every commercial flight is spitting out tons of data, most of which Boeing hadn’t even started processing,” Sharma explains. “We need a lot of infrastructure to process that data, so we can mine useful information to grow the business. So if a part is going bad, the plane can relay that information to the airport where the flight is going to land. Engineers and mechanics would be ready with that part. So time-to-service is lower. The mechanics spend less time maintaining that plane. And the airline wants the plane to be in the air as much as possible. Better data management means better value for our customers.”

Whether addressing information from the air, manufacturing facility, or provide chain, Boeing noticed a rising want to deal with database processing in actual time. The firm’s workloads usually resembled on-line transaction processing (OLTP) datasets, comparable to these utilized by monetary establishments and resolution assist system (DSS) datasets, that are sometimes read-only.  Boeing usually dealt with each sorts concurrently in mixed-workload situations, particularly when combining with different datasets coming in from provide chain companions.

One of Boeing’s database tasks concerned an operational information retailer (ODS) working Oracle actual utility clusters (RACs) on commodity methods. Including dev and check methods and the manufacturing surroundings, the ODS infrastructure consumed roughly 100 servers. Each cluster server ran OLTP and DSS workloads with 1TB of memory and greater than 1PB of combination storage. As Sharma describes it, cluster efficiency was “not at par”, and the {hardware} had reached its finish of life.

New tech for brand new potentialities

Boeing despatched out a number of RFPs and finally examined  three choices: an HPE answer, commodity {hardware} primarily based on the latest-generation Intel processors, and Oracle’s Exadata Database Machine X8M. One of the latter’s main benefits, Sharma says, was its incorporation of 1.5TB of Optane persistent memory for every server. Optane’s DRAM-class latency and higher-than-DRAM capability factors made it a super match for Oracle’s OLTP-type workloads, which profit from very quick entry to small chunks of information from cached storage. (In the X8M’s structure, persistent memory occupies the primary of three routinely managed storage tiers, adopted by NVMe-attached NAND and eventually exhausting disk storage.)

  • Join Transform 2021 for a very powerful themes in enterprise AI & Data. Learn extra.

Boeing’s industrial operations gathers all the corporate’s manufacturing and provide chain information. It’s then aggregated into an information retailer, the place manufacturing facility ground transactions are built-in with provide chain information. Boeing makes use of the outcomes of these integrations to make knowledgeable choices on what stock is required for every particular airplane.

“We need to get data from storage faster and keep it closer to compute,” explains Sharma. “How much needed data can reside in memory? When data is closer to compute, how fast can I process it? With the introduction of persistent memory and 100 gigabit-per-second RDMA over converged Ethernet (RoCE) network fabric — which lets nodes request data directly from PMem rather than going through the entire stack —  we saw an opportunity to eliminate most of the latency.”

To be clear, Optane “PMem” doesn’t substitute DRAM. Rather, it serves as high-capacity risky system memory whereas bumping DRAM right into a high-speed caching position. Or it might probably act as a super-fast, non-volatile (persistent) cache for storage. Boeing primarily employs the latter, by way of Oracle’s information accelerator performance.

However, Boeing has a second use in “persisting” (saving into non-volatile storage) Oracle redo logs —  a mandatory step earlier than a transaction will be dedicated. Redo logs usually get endured to typical SAN storage, which introduces substantial latency. That step beforehand accounted for lots of Boeing’s lag, particularly because the group’s redo lot sizes common round 24GB. Trying to persist that quantity of information ceaselessly provides appreciable course of delays.

Thanks to Intel’s App Direct Mode for Optane persistent memory, and Oracle’s current assist for App Direct in its platform, Boeing may deal with each risky and chronic fashions concurrently, says Sharma.

From deployment to outcomes

After months of intensive testing, Boeing deployed its new Exadata servers into manufacturing in June, 2020. Teams consolidated practically 100 commodity servers down to solely two Exadata racks with eight servers every, and a pair of half racks divided throughout two information facilities.

“Overall, based on how much value we could get from any of the options, the Exadata with persistent memory stood out most,” says Sharma. “It was integrated with other Oracle internals, like Oracle Linux and GoldenGate, that we use heavily to bring in data from our OLTP environment.”

According to Sharma, Boeing encountered no points throughout deployment or perhaps a want to adapt its software program to accommodate persistent memory, as Oracle had already accomplished this work with Intel. The solely additional labor arose from Boeing’s coverage in opposition to third-party racks into its information facilities. As a consequence, Oracle had to re-rack its Exadata methods into Boeing’s personal racks over the course of a number of days.

In the months since preliminary deployment, Boeing studies 2x to 10x productiveness positive aspects by switching to the Exadata X8M platform. The greatest database operations enhancements over the earlier  commodity infrastructure got here from bolstered redo log efficiency, Sharma says, including that different examples abound.

“When we run our batch processing, multiple jobs run overnight. Workloads that consistently used to take 14 hours now take about two hours. This really matters because of various work shifts coordinating across time zones. It becomes very challenging when a shift starts and needs results from another group that hasn’t finished. With jobs finishing faster, our shifts can make better decisions.”

Going ahead: Doing extra with fewer sources

Despite its appreciable server consolidation, Boeing says it nonetheless has ample capability left in its new Exadata answer. This opens the door to taking up extra workloads from different duties or teams. Sharma expects container know-how to play a task in additional consolidation, permitting engineers to cleanly separate, say, manufacturing workloads from the provision chain, engineering, or analytics. Containers may additionally assist with compliance in Boeing’s authorities operations, he provides.

Beyond consolidation and information isolation, the corporate says it might probably now preserve and handle workloads with fewer sources. For instance, reasonably than needing separate directors  to handle completely different layers of the answer (storage networking, and many others.), one consolidated staff now can handle your complete stack. Sharma says this turns into doubly essential as a result of Boeing’s Oracle databases run on a mixture of Linux, IBM AIX, HP-UX, and different working methods. Having one normal platform reduces  spending on sources and infrastructure. Again, it’s concerning the efficiencies of consolidation.

“There has always been a race between the different components of the infrastructure stack,” Sharma notes. “These advances in compute and persistent memory allow customers like us to process more data in a timely fashion. Data is exploding, and so is the demand for storing, retrieving, and processing the data set. These innovations will give us more leverage in consolidating workloads and doing more data analytics locally. Especially with container technology onboard, we can bring in petabytes of data for processing in one location and help drive the business.”

VentureBeat

VentureBeat’s mission is to be a digital city sq. for technical decision-makers to achieve information about transformative know-how and transact.

Our web site delivers important info on information applied sciences and techniques to information you as you lead your organizations. We invite you to change into a member of our neighborhood, to entry:

  • up-to-date info on the themes of curiosity to you
  • our newsletters
  • gated thought-leader content material and discounted entry to our prized occasions, comparable to Transform 2021: Learn More
  • networking options, and extra

Become a member