1 Unleashing the Power of Big DataEric Greffier Director, Business Solutions & Expertise 23-November 2016

2 Disrupt or Be Disrupted41% of executives say digital disruption increases risk of being put out of business Digital Vortex 2015 “New technologies combine to create a business innovation platform, not just a technology platform, helping transform every industry on the planet.” IDC (Dec. 2014) Today I will connect the dots between Disruption, Digitization, Innovation, and the Partner Ecosystem Disruption is happening everywhere. It’s happening in every industry, in every country, in every company. This past summer I was at customer event with 150 CIOs in room. I was part of a panel and there was a SVP from IDC who was facilitating the panel. He kicked of the entire session looking right at those CIOs saying just that. He said, “Disruption will happen in every one of your companies.” Now, for the CIOs in the room, that’s kind of scary. However, for everyone in this room that is pretty darn good news. [CLICK: animation/build comes up] Because the answer to disruption is digitization. Digitization is combining new technologies in new ways to help a customer either transform their operating model or transform their business model. So, in other words it is about helping them engage their customers very differently. Helping them engage with their employees in new ways to increase productivity. Or help them to automate and find new business processes that they can introduce. So, in other words, the other reason this is really good for all of us is that NOW all of a sudden the Line of Business buyers who are watching what their competitors are doing with technology, they realize that technology now is going to help create their new ways to differentiate and complete. We are already seeing this put into action at Cisco with visitors to our Customer Experience Center. Chuck Robbins, our CEO, recently shared that in the last two quarters 70% of visitors have been executives from outside IT. We’ve had thousands of visitors so this change is dramatic and reflects how customers are looking at digitization to enable, differentiate, and define their business strategies.

3 Analytics & Automation Software…Driving Insights to ActionAnalyze Access Engage Automate Edge Datacenter Cloud Distributed Analysis App Services Vendor & Partner Ecosystems Sense… …Act Citizens Customers Employees Devices Data Warehouses Apps Big Data Stores Virtual Aggregation Contextual Interactions This digital transformation requires a new approach to data, analytics and action. It requires that you sense what is happening across the network and orchestrate timely and informed action. It is a model that requires analytics and automation platforms built for today’s hyper connected world - platforms that enable: Agile access to data of interest everywhere by virtually aggregating and query data living across the network as if it were in a single place - integrating hundreds of data sources in days instead of months and arming more people with the critical data they need, in the way they need it. You also need to analyze this highly distributed data, that lives in the datacenter, the cloud, and more and more at the edge of the network, wherever it makes the most sense – based on the business requirements, technical feasibility and cost implications. You need to be able to automate and orchestrate process and services interactions across your business and your vendor and partner ecosystem And finally, you need to deliver contextual interactions. Improving the digital experience for customers, employees, and citizens is a top priority for organizations , but delivering personalized and engaging interactions requires processing information about who and where a person is to deliver contextual insights at the right time, in the right location, and how that person wants to consume the information.

4 The Forrester Wave™: Big Data Hadoop-Optimized Systems, Q2 2016"Cisco Systems provides a viable midsized system at attractive price point. Cisco UCS Integrated Infrastructure for Big Data provides a secure and scalable infrastructure to support enterprise requirements. Cisco’s UCS solution comes pretested and prevalidated for Cloudera, Hortonworks, IBM, and MapR, providing a lower-cost and scalable storage platform to support Hadoop deployments. Management tools such as Cisco UCS Manager and Cisco UCS Director allow for simple configuration of big data Hadoop clusters that can adapt dynamically to changing workloads. Cisco’s key differentiators lie in its ability to offer a wide range of configurations, its strong focus on internet-of-things (IoT) use cases, and its broad partner ecosystem.” Recognition for Cisco UCS continues! Forrester Research recently positioned Cisco as a “Leader” in their Forrester Wave™: Big Data Hadoop-Optimized Systems, Q report.

5 A Broad Perspective To Set The SceneAPP SaaS PaaS IaaS APP Data Center Cloud Edge / IoT That opportunity lies in finding ways to disrupt or run the risk of being disrupted. However, for you to be able to stay ahead of disruption a broader perspective is required one that goes beyond the traditional data center model. We must now go beyond the data center, expanding its value into the cloud, as well as the edge and IoT. It’s a view not limited to just a network topology or architecture discussion, but rather, it is about the new business models made available to you by the cloud. And when it comes to the edge and IoT, we only have to look to the market to see where the opportunity lies. When we look at the true disruptors in the market, whether they be Uber, Spotify, Airbnb, or Stripe, they engage their customers at the edge of the network with smart applications that leverage disruptive technologies, data, mobility, and devices. But going beyond the data center also requires extending security. Organizations are under unrelenting attack, with security breaches happening every day from a global community of attackers targeting organizations of all sizes through a network perimeter that has large dissolved due to trends like mobility and BYOD. Security from the data center to the edge requires the protection against a wider array of attack vectors. To digitize your business processes and realize the IoE $19 trillion opportunity in value by 20201, you need to take advantage of this expanded landscape to drive business outcomes, thus requiring you to evolve architectures for infrastructure, development, and security, as well as business and governance models Additional Notes: The edge is defined as the IT boundary closest to the user.  In the past, this has been limited to the branch office.  While the branch (or retail store) is still critically important, the Internet of Things is changing that definition.  Now, the edge is often the IT boundary closest to the sensors, customers, security cameras and other devices that are feeding and processing information for either immediate use or analysis later. According to the Cisco 2015 Annual Security Report, malware is becoming increasingly sophisticated and elusive; cybercriminals are launching attacks through a variety of attack vectors including tools that users trust or viewed as benign; and targeted attacks are on the rise, creating a persistent, hidden presence within an organization and from which to execute their mission. Since 2009, we have seen a 66 percent compound annual growth rate of detected incidents. ________________________________________ 1 IoE research: © Cisco and/or its affiliates. All rights reserved. Cisco Confidential 5

6 The Explosion of Unstructured DataValue 2005 2015 2010 More than 90% is unstructured data Approx. 500 quadrillion files Quantity doubles every 2 years Most unstructured data is neither stored nor analyzed! 1.8 trillion gigabytes of data was created in 2011… 10,000 GB of Data (IN BILLIONS) STRUCTURED DATA UNSTRUCTURED DATA Source: Cloudera

7 A New Approach is Needed to Reach and Analyze That DataDigital Enterprises Create Data at the Edge A New Approach is Needed to Reach and Analyze That Data Widely Distributed, Streaming, Short Shelf Life, Too Big to Move Structured Data Unstructured Data Data Streaming at the Edge “Most data will be processed at the edge” (mobile devices, appliances, routers) 37% Traditional Data Warehouse Big Data Store REAL-TIME LONG-TERM SHORT-TERM Make a Decision NOW Make a Decision Today Plan for a Business Decision Three years from now, where will most data generated by IoT solutions be processed?

8 Choose Your Database Based on Your Use-caseBig Data High Stream-Analytics Real-Time IoT Analytics < ms Response Time Complex Event Processing ms Massively Parallel (MPP) Real-Time In-Memory DB 1 sec Operations Analytics Batch-Analytics Doc Store 1 min OLAP 10 min OLTP Reporting 1 hr Low 4 hrs Gigabyte Terabyte Petabyte

9 Delivery Model - Solution Building Blocks for Big Data & AnalyticsCisco Connected Analytics Collaboration Analytics Contact Center Location Analytics Network Analytics Analytics Ready Infrastructure IT Operations IT Security Business Analytics TS, AS and CCS Analytics/ Business Intelligence Data Integration Virtualization Preparation ETL PARTNER Integrator / Data Science In-Memory MPP/NoSQL Data Management Cisco Integrated Infrastructure Networking Servers & Storage Security Management So let’s take a look at what solution components we can sell, and how our customers will build their Big Data solutions. This graphic shows you the various blocks you can use to build an end-to-end solution, or you can simply just focus on an individual block. Let’s start at the bottom and move our way up. Cisco Integrated Infrastructure – If you open up the metaphorical ‘file cabinets’ of the database world, you’ll discover servers, storage, networking, management and hopefully security. We naturally do all of this, and we do it very, very well. A typical Hadoop solution has servers with direct attached storage over a distributed computing environment… servers with storage built into the server, keep the data as close as possible to the compute to reduce latency and increase processing performance. Data Management – We have three options: In Memory databases like SAP Hana, Hadoop vendors such as Cloudera, MAPR and Hortonworks (which are on our GPL) and MPP Databases such as Actian, and NoSql solutions such as Mongo DB and Oracle NoSQL. While our infrastructure supports all three variants, the leading solution for Big Data is Hadoop, and justifiably receives most of our attention and focus. Hadoop on UCS is a validated Cisco design and our performance is exemplary. Data integration – is the layer linking the data management system to all the various sources of data and to Analytics software systems. This is a critical layer where our trump card is our Cisco Data Virtualisation software (from the Composite acquisition). Imagine you have a traditional data warehouse alongside the more cost effective Hadoop data warehouse. How do you create a single view of all of your data before it gets fed into the analytics software? Cisco Data Virtualisation is the answer to this – it unifies or federates a single view into ALL of your data sources. This provides a massive competitive advantage for Cisco and is our key differentiator. It has a massive impact on performance and effectively ‘unlocks’ ALL the data for use in the analytics software. Analytics/business intelligence – Let’s first look at our Analytics Ready Infrastructure – this is the platform to run 3rd Party analytics offers… it expands our opportunity tremendously in the accounts as these analytics providers require a high performing infrastructure to make their software work… a symbiotic relationship in its truest sense. The second area is our own Analytics Solutions. Our new Cisco Connected Analytics solutions provides great insight and recommendations to our customers to manage their Cisco technology investments better. This establishes us a key partner in the management and decision making of our customers’ IT operations. Security, TS, AS and CCS – Make sure the platform is secure in all of the blocks and wrap services around our solutions. We also have experts in consulting to advise our clients on their data strategy. PARTNER – our Channel Model supports all stages of a project, from the first step (Consulting), the data model (Data Science) to the integration (Systems Integrator).

10 Cisco Integrated Infrastructure for SAP HANAHANA Appliance HANA Compute / Network 4-16 x 2TB B460 HANA nodes 2x 6332 FI 1x N3K (or N9K / ACI) Mandatory HWCCT Appliance KPI met Mandatory Solution Support Opt-out Cisco Managed Service HANA Storage Scalable C240 storage 3-8x C240 Storage Nodes 24x 1.8T HDD RHEL 6 MapR Data Platform Cisco Integrated Integrated for SAP Hana (Appliance model) Main use case is for analytics work loads What’s new includes: 40G network speed C240 servers with MapR cluster storage, allows scaling in smaller storage chunks – MapR converged data platform We used MapR because there is proprietary IP in their converged data platform that is not included in the other hadoop vendors and that provides for the enterprise storage capabilities. Services are mandatory – Cisco Managed Services where Cisco will fully manage your solution and the customer has the option to opt out. But in so doing then the customer recognizes that the customer is in charge of all the patches and updates Note to PSS: SAP Certifies appliances prior to TDI mode, so configurations on slide 18 and 19 have different storage options (Appliance was a prerequisite certification to TDI which allows traditional enterprise storage

11 Cisco is a Leader in Big Data InfrastructureExternal Benchmarks Industry's Best TPCx-HS Performance in ALL Categories IDC Key Performance Indicators for Cisco UCS Integrated Infrastructure for Big Data 32% Lower Operational Cost 56% Faster Time to Market 26% Data Scientist/Analyst Productivity Gained Forrester The Forrester Wave(TM): Big Data Hadoop-Optimized Systems, Q2 2016 A Leader in Hadoop Optimized Systems Better Cheaper 1TB 11% 30% 3TB 14% 10% 10TB 28% 3% 30TB 12% 33% 100TB The Best— No Comparison “Cisco's key differentiators lie in its ability to offer a wide range of configurations, its strong focus on internet-of-things (IoT) use cases, and its broad partner ecosystem.”

12 Big Data Projects Requires Right Networks

13 Compute/Storage LanguageHow? Applications Are Now Everywhere Teach the Infrastructure the Language of the Applications Policy-Driven Infrastructure Application Language Application-tier policy and dependencies Security requirements Service-level agreement Application performance Compliance Geo dependencies Network Language Compute/Storage Language Security Language

14 Application Centric InfrastructurePolicy Driven APIC Application Open Manageable & Visible Cost Effective Simple Secure Scalable Physical Virtual Multi-Hypervisor Security Services $ ACI NEW systems ARCHITECURE FOR THE DC. DESIGNED FOR THE DIGITAL ERA COMPLETE CHANGE OF THINKING – we built this for applications MOVE TO FABRIC ARCHITECTURES New form of network substrate (leaf spine) (1) We build a better network - leaf spine, scalable, predictable                  scale out - move intelligence from central to edge (leaf) (2) SERVERS AND SERVICES ARE AT EDGE (MODERNISE) (3) APIC – WE INDUSTRIALISE THROUGH …APIC SIMPLE, - ONE NETWORK, NOT 2. simple to operate. SELF CONFIGURING NETWORKING, SCALABLE – LEAF SPINE, 40G. DECLARATIVE MODEL SCALES BETTER SECURE – any vendors firewalls, multi-tenant, abstract away from VLANS. STATELESS FIREWALL. COMPLIANCE. SECURITY ZONES. SELF DOCUMENTING NETWORK SECURITY - THE MORE WE DO, THE LESS IT IS. NETWORK. WE MUST DOCUMENT THE POLICY BEFORE WE DEPLOY. MANAEGEABLE – all devices from a single pane of glass. VISIBLE – health scores. WE HAVE CO-ORDINATION BETWEEN THE OVERLAY AND THE UNDERLAY BUT MOST IMPORTANT (OVER ARCHING COMMENT) – THE DATA CENTRE IS A VERY DIVERSE SPACESEGUE: BUT I SAID THAT IN SOFTWARE DEFINED NETWORKS, THE VALUE IS IN THE DEFINITION, LETS LOOK AT THIS IN A BIT MORE DETAIL

15

16 Cisco UCS Integrated Infrastructure for Big Data 3rd generation of Cisco UCS Common Platform Architecture Pre-tested, pre-validated and documented best practice designs optimized for performance and capacity – lowering risk and TCO Designed to scale from small to very large as business demands Unified and centralized management with seamless Integration with enterprise applications Easy to {order, deploy, service} Single SKU Big Data Bundles High Performance Big Data Extreme Capacity Capacity Optimized Performance Optimized UCS 6300 Series Fabric Interconnects LAN, SAN, Management UCS C220/C240 M4 Servers UCS C3260