The tech talk was given by Jim Walsh, Salesforce SVP Infrastructure Engineering in May 2017.
The presentation provides a brief overview of Salesforce Cloud Infrastructure and Challenges.
Salesforce Cloud Infrastructure and Challenges - A Brief Overview
1. Jim Walsh
SVP, Infrastructure Engineering
Storage Cloud
jimw@salesforce.com
@jim1walsh
Salesforce
Infrastructure Engineering
A Brief Overview
May 4, 2017
2.
3. Forward-Looking Statements
Statement under the Private Securities Litigation Reform Act of 1995:
This presentation may contain forward-looking statements that involve risks, uncertainties, and assumptions. If any such uncertainties materialize or if any
of the assumptions proves incorrect, the results of salesforce.com, inc. could differ materially from the results expressed or implied by the forward-looking
statements we make. All statements other than statements of historical fact could be deemed forward-looking, including any projections of product or
service availability, subscriber growth, earnings, revenues, or other financial items and any statements regarding strategies or plans of management for
future operations, statements of belief, any statements concerning new, planned, or upgraded services or technology developments and customer contracts
or use of our services.
The risks and uncertainties referred to above include – but are not limited to – risks associated with developing and delivering new functionality for our
service, new products and services, our new business model, our past operating losses, possible fluctuations in our operating results and rate of growth,
interruptions or delays in our Web hosting, breach of our security measures, the outcome of any litigation, risks associated with completed and any possible
mergers and acquisitions, the immature market in which we operate, our relatively limited operating history, our ability to expand, retain, and motivate our
employees and manage our growth, new releases of our service and successful customer deployment, our limited history reselling non-salesforce.com
products, and utilization and selling to larger enterprise customers. Further information on potential factors that could affect the financial results of
salesforce.com, inc. is included in our annual report on Form 10-K for the most recent fiscal year and in our quarterly report on Form 10-Q for the most
recent fiscal quarter. These documents and others containing important disclosures are available on the SEC Filings section of the Investor Information
section of our Web site.
Any unreleased services or features referenced in this or other presentations, press releases or public statements are not currently available and may not
be delivered on time or at all. Customers who purchase our services should make the purchase decisions based upon features that are currently available.
Salesforce.com, inc. assumes no obligation and does not intend to update these forward-looking statements.
4. Trusted Customer Success is our #1 Value
Delivering the highest standards in system availability,
performance and security is our top priority.
5. 118% ↑
124% ↑
100% ↑
104% ↑
94% ↑
Enabling Phenomenal Customer Success
99.97% Availability
50 Production Instances
490 Billion Transactions
• 230ms average latency
10 Data Centers
• 1st in EMEA – London
194 MC Customer Databases
247 Billion emails sent
99.98% Availability
109 Production Instances
1.1 Trillion Transactions
• 210ms average latency
20 Data centers
• 3 in EMEA – London, Paris, Frankfurt
395 MC Customers Databases
478 Billion emails sent
Growth Across The Clouds
2014 2016
6. Infrastructure Engineering Vision
Continuous Innovation
Technology
• Upgrades across entire infrastructure
Process
• Leverage new technologies to streamline and
simplify
People
• Do, Measure, Learn, Train, Feedback
Always Improving Availability, Performance and Security
Trusted Multitenant Infrastructure
7. Infrastructure Engineering Principles
One Salesforce
• Leverage best practices and common processes
Design Principles
• Service Ownership: Software Engineers operate
the services they create
• Recovery-oriented software architectures
• Prefer scale-out architectures
• Simple, consistent hardware
• Measure and improve key metrics such as
availability through continuous application and
platform changes
Integrating and Scaling
8. Software Defined Everything
• Compute
• Storage
• Security
• Networks
Task
Manual
(“Ops”)
Automated
(“DevOps”)
Autonomous
(“No Ops”)
Sets the goal Human Human Human
Decides when to start the work Human Human Machine
Adjudicates work priorities Human Human Machine
Does the work Human Machine Machine
Generates the validation report Human Machine Machine
Interprets the validation report Human Human Machine
Handles failures Human Human Machine
Handles exceptions Human Human Human
Evolving for Scale
From Manual & Automated to Autonomous Operations
9. Business Continuity For All Salesforce Customers
Large or small, global business has no downtime
Faster Recovery
From Events
Reducing
planned/unplanned
downtime
Faster
Transactions
Fast Global
Performance
11. Every instance (NA,EU,AP)
has full-capacity secondary
• Including Sandboxes (CS)
Each data center has active
instances
All instances level services
are redundant
• Scaled as necessary for
resilience and performance
All shared services are
scaled to run at full capacity
in a single data center
Availability and Performance Strategy
Fast, Regional Site Switching
Primary Instance
Secondary Instance
Dallas
NA7 CS50NA34 CS3
CS59
NA33 CS30NA3 NA45
NA8 NA39 CS51CS12
Phoenix
CS52NA32 NA44 CS13
NA7 CS50NA34 CS3
CS59
NA33 CS30NA3 NA45
NA8 NA39 CS51CS12
CS52NA32 NA44 CS13
Group 1 Group 1
Group 1 Shared Services Group 1 Shared Services
Data Center Shared Services Data Center Shared Services
Salesforce Confidential
12. Technology
• Oracle 11g + Active Data
Guard
Process
• DB active in read-only mode
• Reduced process steps
People
• Regular training and execution
of process
Result
• 4x improvement in recovery
time
Continuous Enhancement for Availability
DB Log-level Replication Within Region
Encrypted
Async DB
Replication
Data Guard
Replication
Primary Instance
Application
Servers
Production
DB Cluster
Standby
DB Cluster
Secondary Instance
Data Guard
Replication
Application
Servers
Production
DB Cluster
Standby
DB Cluster
19. Keystone
Catalog
Keystone
Catalog
Standardizing the API: Org and Location Namespaces
Primary Datacenter Secondary Datacenter
Fileforce Servers Fileforce ServersBackup Backup
Org Namespace
Location Namespace
New Store New Store
20. Some Challenges in Large Scale software-defined Storage
❑ Leveraging Immutability in building Distributed Stores
❑ “Immutability changes everything” – how to do consistency, load sharing, backup, … ?
❑ Caching and Tiering Strategies for Distributed Object Stores
❑ How much of caching and tiering decisions need to be local (within a OSD) and global?
❑ Resource Management and Orchestration for Stateful Containers
❑ Reducing tail latencies to a distributed store
❑ Especially sequential writes – log and LSM workloads
❑ …
21. Tips for Success
• Build it!
• Be it tools or systems – create artifacts!
• Use Open Source where possible: easier to share, get feedback and get it used by others
• So many are already out there - Kubernetes, Ceph, Hbase, BookKeeper ….
• Start one, if necessary!