Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Driving Towards Cloud 2015: A Technology Vision to Meet the Demands of Cloud Computing Tomorrow
1. Driving Towards Cloud 2015:
A Technology Vision to Meet the Demands
of Cloud Computing Tomorrow
Steve Pawlowski
Intel Senior Fellow
CTO, Datacenter and Connected Systems Group
GM, Datacenter and Connected Systems Pathfinding
2. Cloud Computing Today
On Anything
More Visual Content
With More Flexibility
On the Move
Delivered With Unanswered
Independent of time Questions –
or place • Secure?
• Reliable & Available?
• Standards Ready?
From Some Place That
• If yes, who ensures it? I Don’t Have to Manage
But Connects Me to Everything
3. Security and Privacy Is the
Biggest Obstacle
High Risk Threats
• Well resourced, motivated cyber warriors
• Do you know where
• Aggressive and Pervasive your data is?
• Referred to as Advanced Persistent Threat
• Who can see it?
Medium Risk Threats
• Criminals to steal identity and money • Who can modify it
• Varying technical sophistication without a trace?
• Who can aggregate,
Low Risk Threats summarize, and
• Internet Pollution embed it for purposes
• Threats against every user other than yours?
Threat Categorization from ISIMC Guidelines for Secure Use of Cloud Computing V0:10
4. Energy Efficiency Is A Big Issue
2006: Data Centers
account for 1.5% of
total US Energy
Consumption
US Environmental Protection Agency Report on
Server and Data Center Energy Efficiency – Aug 2007
Efficient Data Center Energy Use Power Breakdown in a Server Platform
5. Think About Energy Efficiency
Differently
Grid
Distribution Water / On Site Energy
Carbon Generation Storage and
Facility (WUE/ CUE) Recycling
PUE Rack and Row Level Infrastructure
Optimization Infrastructure Efficiency
(PUE)
Server Power Power
Supply Managed Proportional Autonomous
Servers/ Data Center Power Aware
Fans and Cooling
Racks
Motherboard/VRs
Package
Energy Efficient Compute, Network, and Storage
Die
Gates 2010 TODAY 2015 2020
Sizes are illustrative
Focus All the Way From “Grid to Gate”
7. Surge In Devices, Users, and
Social Networking
Mobile Internet
2.5 Billion
Connected
Users
Devices/Users
by 2015
>10 Billion
Connected
Devices By
2015
Source: Gérald Santucci, Head of Unit European Commission DG INFSO 20 Years of ICT Revolution:
The Unceasing ‘Grail Quest’ , March 2009
Cisco Visual Networking Index
8. Surge in Storage
Storage Capacity Shipped
“More video uploaded to
YouTube in the past 2
670%
months than if ABC, CBS
and NBC had been airing Growth
new content 24/365 since
1948.”
Exabytes
— Gartner
“Big Data last 5 years:
800% growth; 80%
unstructured effective
tiering needed.”
— Gartner
9. Surge in Sensors
• Downloaded Apps Require Sensors
• ~1B GPS, ~900M Accel, ~600M Compass,
~450M Gyros by 2014
10. Surge in Video Traffic
Mobile Traffic Today 2015 Mobile Traffic
40% 66%
~40x
Video Video
90 PB/month
7M paid video
subscribers
3600 PB/month
700M paid video subscribers
Source: Gérald Santucci, Head of Unit European Commission DG INFSO 20 Years of ICT Revolution:
The Unceasing ‘Grail Quest’ , March 2009
Cisco Visual Networking Index
11. Along With the Cloud’s Key
Characteristics
Elasticity
Self Service
Pay As You Go
12. So, What are the
Key Technical Areas
to Focus On?
13. Creating a Strong Foundation For
Cloud Computing
Match WL With Platform Big Data Client Aware
and Automate and Analytics Computing
No “The” WL in the Cloud Future Computing Not Just Centralized
Embrace Mixes of Different Platforms dominated by Computing
specialized for different classes of continuous
applications ingest, Adaptively distribute
integration, and functionality among
Automate to maximize efficiency & analytics of large,
robustness as cloud scales – across HW, DC, Local Servers, and
growing, and live Edge Devices
SW, and Problem Diagnosis data
While Keeping A Focus On Security At All Times
14. Match Workload With the Platform
Integrate
Embrace Heterogeneity New Memory Technologies
Explore & extend range of Explore how cloud apps can
apps for platforms by exploit new NVM
overcoming OS limits, technologies such as
memory limits, and Restive Memories to
scalability issues in overcome memory scaling
specialized platforms of issues
heterogeneous nodes
15. The Focus Areas of Automation
Resource Scheduling Software Correctness
& Task Placement & Productivity Problem Diagnosis
Devise mechanisms Automated tools for Explore new
and policies to software upgrade techniques for
maximize energy management, diagnosing problems
efficiency, runtime correctness at cloud complexity
interference checking, and and scale
avoidance, data programmer
availability and productivity
locality
16. Develop Methods to Quantify
and Measure Usage
100% • Customer workloads vary in
90% their infrastructure
80% demands: Typically:
70%
Memory Utilization, Storage
60%
50% IO, Network Throughput
40%
30% • Current measurement
20% metric for Cloud Providers –
10%
Number of Systems and
0%
Time Used
• The Future: HW/SW co-
designed solutions to
monitor data for better
My billing System billing Libraries billing
scheduling & metering
17. So, What Does Big Data Look Like?
An Example: The Semantic Web
Main Stream Web
Links to Resource Links to Web of Data
Resource Resource
Links to Links to
Resource
Computers Understand the Meaning
Resource Resource of Data On the Web – See the “Big
Links to
Picture”
Semantic Web
Requires Graph Databases for
Has Manual Software Requires Geospatial and Temporal Search of
Document Library
Semantic Stream
Author
Person
Address Image
Evolving New Topics and Linguistic
Place Image Differences Make This Challenging
Big Data is - Diverse, No Schema, Uncurated,
Inconsistent Syntax and Semantics1
1 Amplab – Dr. Patterson
18. Towards Efficient Frameworks for
Big Data
14
• Popular current big data
12
Slowdown Factor
frameworks very inefficient
10
– 3-13x slower after
8
optimization
6
• Bunch of reasons why 4
– Artifacts of implementation 2
– Costs for desired features 0
– Unexpected I/O effects MapReduce MapReduce Hadoop MapReduce Hadoop
Grep TeraSort TeraSort PetaSort PetaSort
Frameworks for Advanced Characterization & Better
Machine Learning Algorithms Programming for Big Data Apps
Focus on selective iteration and Focus on areas such as processing of
exploitation of dependencies among data from the web, social network
data-items in the data set being interactions, malware analysis, video
analyzed image processing, and HPC apps
19. The Storage Challenges With Big Data
Persistent Data
What needs to be stored and for how long?
Data Organization
• Many-to-One Network Issue:
Multiple simultaneous reads to pick up every stripe
• Energy Efficiency Issues:
Servers need to be powered on all the time
20. Clients Must Be Cloud Aware and
Vice Versa
• Edge devices will be directly involved in
many “cloud” activities
–May even form clouds of their own
• Bridge to physical world: Edge devices bring
sensors, actuators, “context” to the cloud
Cloud Assisted Mobile Wide Area Replication,
Client Computations Consistency, and Deduplication
Enable applications whose execution Address edge connectivity issues by
spans client devices, local servers, creating new ways to eliminate wasted
and the cloud bandwidth and reduce latency
21. Balancing Compute & Client
Side Analytics
18,000 3,000,000
16,000
GB Transfer/Day
Monitored Cores
2,500,000
14,000
12,000 2,000,000
10,000
1,500,000
8,000
6,000 1,000,000
4,000
500,000
2,000
0 0
2011 2012 2013 2014 2015 2016
Total Devices Daily Raw Data Daily Data Transfer
Transfer (GB) w/Client Side Analytics (GB)
Source: Intel Cloud Builders Guide
• Leveraging local resources • Monitored Cores and monitored
reduces the data center data from servers in data centers
workload and associated increases at a steady pace
network traffic • Significant analytics required to
• Increases the VDI scalability manage, make sense, and take
and enhanced quality of service action
for end users • Client Side Analytics – the only
alternative
22. More Unknown Than Known Out There
Predict – But Find a Way to Recover
Prevent the Security & Trust Infractions Prevent
Detect
Detect the Security & Trust Infractions Miss
Recover
Miss
Recover quickly and gracefully from the Security & Trust
Infractions
23. The Future is a World of Many
Clouds
• The Cloud – while still evolving – enables
many possibilities
• Challenges in Cloud Computing has so far
mostly kept it to social networks; Security
and Privacy tops the list
• Three Key Focus Areas:
1) Match WL with the Platform & Automate
2) Big Data & Analytics
3) Client Aware Computing
26. Risk Factors
The above statements and any others in this document that refer to plans and expectations for the second quarter, the year and the
future are forward-looking statements that involve a number of risks and uncertainties. Words such as “anticipates,” “expects,”
“intends,” “plans,” “believes,” “seeks,” “estimates,” “may,” “will,” “should,” and their variations identify forward-looking statements.
Statements that refer to or are based on projections, uncertain events or assumptions also identify forward-looking statements. Many
factors could affect Intel’s actual results, and variances from Intel’s current expectations regarding such factors could cause actual
results to differ materially from those expressed in these forward-looking statements. Intel presently considers the following to be
the important factors that could cause actual results to differ materially from the company’s expectations. Demand could be different
from Intel's expectations due to factors including changes in business and economic conditions, including supply constraints and
other disruptions affecting customers; customer acceptance of Intel’s and competitors’ products; changes in customer order patterns
including order cancellations; and changes in the level of inventory at customers. Potential disruptions in the high technology supply
chain resulting from the recent disaster in Japan could cause customer demand to be different from Intel’s expectations. Intel
operates in intensely competitive industries that are characterized by a high percentage of costs that are fixed or difficult to reduce in
the short term and product demand that is highly variable and difficult to forecast. Revenue and the gross margin percentage are
affected by the timing of Intel product introductions and the demand for and market acceptance of Intel's products; actions taken by
Intel's competitors, including product offerings and introductions, marketing programs and pricing pressures and Intel’s response to
such actions; and Intel’s ability to respond quickly to technological developments and to incorporate new features into its products.
The gross margin percentage could vary significantly from expectations based on capacity utilization; variations in inventory
valuation, including variations related to the timing of qualifying products for sale; changes in revenue levels; product mix and
pricing; the timing and execution of the manufacturing ramp and associated costs; start-up costs; excess or obsolete inventory;
changes in unit costs; defects or disruptions in the supply of materials or resources; product manufacturing quality/yields; and
impairments of long-lived assets, including manufacturing, assembly/test and intangible assets. Expenses, particularly certain
marketing and compensation expenses, as well as restructuring and asset impairment charges, vary depending on the level of
demand for Intel's products and the level of revenue and profits. The majority of Intel’s non-marketable equity investment portfolio
balance is concentrated in companies in the flash memory market segment, and declines in this market segment or changes in
management’s plans with respect to Intel’s investments in this market segment could result in significant impairment charges,
impacting restructuring charges as well as gains/losses on equity investments and interest and other. Intel's results could be affected
by adverse economic, social, political and physical/infrastructure conditions in countries where Intel, its customers or its suppliers
operate, including military conflict and other security risks, natural disasters, infrastructure disruptions, health concerns and
fluctuations in currency exchange rates. Intel’s results could be affected by the timing of closing of acquisitions and divestitures.
Intel's results could be affected by adverse effects associated with product defects and errata (deviations from published
specifications), and by litigation or regulatory matters involving intellectual property, stockholder, consumer, antitrust and other
issues, such as the litigation and regulatory matters described in Intel's SEC reports. An unfavorable ruling could include monetary
damages or an injunction prohibiting us from manufacturing or selling one or more products, precluding particular business practices,
impacting Intel’s ability to design its products, or requiring other remedies such as compulsory licensing of intellectual property. A
detailed discussion of these and other factors that could affect Intel’s results is included in Intel’s SEC filings, including the report on
Form 10-Q for the quarter ended April 2, 2011.
Rev. 5/9/11