01. Sean McKeown, Cisco, Visits theCUBE!. (00:21)
02. Talk About Your Role And How It Relates To Big Data. (01:01)
03. How Do The Requirements Of Big Data Differ From An Architectural Perspective. (02:49)
04. What Is Cisco's Role Here. (04:50)
05. Do You Remove A Lot Of The Complexity. (06:59)
06. Do People Complain About Being Able To Scale To Compute Independant From Storage. (08:39)
07. Discuss ISV's Role. (10:23)
08. What's The Gnarliest Big Data Problem From An Architectural Aspect. (12:10)
09. Do We Need Fast Pipes Between Servers Now. (14:38)
10. What Do You Want People To Know About Cisco And Big Data. (16:12)
#theCUBE #Cisco #BigDataNYC #SiliconANGLE
--- ---
A new frame of mind: How Cisco transforms data architecture | #BigDataNYC
by Timothy Walden | Sep 29, 2016
What’s the gnarliest big data problem facing IT networking in today’s cloud-happy enterprise? For Cisco Systems, Inc., it’s all about hunting down the bottlenecks in the data center. Companies want the most efficient and cost-effective infrastructure, and that requires balance in solutions and price, according to Sean McKeown, Technology Solutions architect, Data Center and Virtualization Sales, at Cisco.
Sitting down with Jeff Frick (@JeffFrick) and Dave Vellante (@dvellante), cohosts of theCUBE, from the SiliconANGLE Media team, during the BigDataNYC 2016 conference, McKeown shared key insights to the networking industry’s response to big data’s recent surge, how it’s impacted data architecture, and where innovation can further the market.
Data Management
Over the past few years some of Cisco’s customers have steadily increased the size of their data clusters, according to McKeown. By utilizing big data solutions that integrate the open source data management platform Hadoop, Cisco is able to provide customers with enough information to run multiple programs on a single datastream.
The key to making sure the data can be processed at the fastest speeds is that Cisco moves the CPU closer to the data and scales horizontally rather than vertically, according to McKeown.
Cisco aims at efficient operations by presenting customers the opportunity to make templates of servers. Customers can establish rules and settings on one server that will apply “global changes across the cluster,” McKeown said.
Gnarly big data problem
Introducing a most interesting topic, Vellante asked McKeown to share the gnarliest big data problem he’s encountered. For Cisco, it’s all about hunting down the bottleneck in IT architecture. It’s easy enough to build a functional system on one’s laptop or a five-node cluster, and even get some good results, McKeown explained. Things get hairy, however, when scaling such utilitarian systems.
“This happens all the time,” said McKeown. “We’ll go to a customer environment where they have done a proof of concept … and build it up, and show this incredible speed up over whatever the previous system is. It’s a very different animal to take that five-server POC and build a 500 server environment. That in itself is a pretty big problem we’re solving reasonably well.”
McKeown went on to share his expectations for emerging technologies, citing cross-point memory and MDME as areas of interest as they become more affordable in the coming years. Harmonizing data architecture means “figuring out how to best take advantage of these technologies at the right price point,” he explained.
The right tools
What makes Cisco so effective at providing high-quality services is its dedication to making sure the customer is prepared for any situation or problem, according to McKeown. As data evolves and increases in volume, the best thing Cisco can do is help customers “figure out how to best take advantage of it all,” he said.
The price equation couldn’t be overemphasized, however, as McKeown explained the importance of an appropriate price point. “I can give you 100GB pipes if you want them, but it’s not going to be cost effective,” he said. “We’ve got to find that right balance.”
#BigDataNYC
#theCUBE
Информация по комментариям в разработке