Download Strategic Directions - Vision for Cloud Operational Analytics

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Recursive InterNetwork Architecture (RINA) wikipedia , lookup

IEEE 1355 wikipedia , lookup

Airborne Networking wikipedia , lookup

Functional Database Model wikipedia , lookup

UniPro protocol stack wikipedia , lookup

Transcript
S T R AT E G I C D I R E C T I O N S
A Vision for Operational Analytics
as the Enabler for Business Focused
Hybrid Cloud Operations
As infrastructure and applications have evolved from legacy to modern technologies with
the evolution of Hybrid Cloud Computing, Network Virtualization, SDDC, SDN, Application
Mobilization, DevOps, Micro-services, enterprises are in the new era of transformation for
infrastructure and application delivery. With this trend, Application and Network Performance, IT service and Security Event Information Management vendors (APM, NPM, ITSM,
SEIM) must evolve their tools to keep up with this paradigm shift. Each of these vendors
has invested significant R&D to collect continuously changing data from various tiers to
deliver operational capabilities. Additionally, enterprises often need to build custom data
collectors and probes where existing market solutions don’t meet their specific operational data needs.
Operational data exists in different forms within an enterprise environment, at different
layers, silos, tools and systems. The variety of data includes raw packet data at the TCP/IP
layer, standard SNMP traps, configuration files, log files, performance and fault data, alerts
and ticketing data, business transaction logs etc. Companies have invested in third party
as well as home grown tools to capture data required to meet their operational needs.
As we follow this trend, we strongly feel that the data itself is becoming commoditized.
There is no shortage of information within the enterprise environment, but making
sense of it to deliver business value is very challenging and complex. The correlation and
analytics of data from disparate sources as well as intuitive visualization to provide just-intime operational intelligence is the pressing need around which value must be created
for the business.
FixStream, Inc.
2680 North First Street
San Jose, CA, USA
www.FixStream.com
© 2015 FixStream Inc.
All Rights Reserved.
As the Cloud Operational Analytics (COA) domain evolves, the core value proposition will
be centered on an intuitive visualization experience, which adapts to various IT, network
as well as business focused end users. COA will support hybrid cloud as this becomes
the de-facto deployment model. It will be built on top of a single logical database that
internalizes activity across multiple data centers and control frameworks. It will be a
highly scalable contextual data correlation and analytics platform. As the touch-point for
end users, the interface will ultimately determine the usefulness of the platform but it will
rely on the quality and completeness of the data it interprets and represents. Therefore,
we believe strongly that COA platforms must have an open and flexible data ingestion
approach to deliver business value.
S T R AT E G I C D I R E C T I O N S
The Open Data Ingestion will enable faster and cheaper collection of data of many
different types and sources across the enterprise ecosystem. It takes care of the complexities of processing the data, but focuses on providing simple interfaces to ingest data in
a scalable way. The platform will provide APIs, SDKs and supporting documentation for
development of mapping logic, AAA policies and runtime execution of the logic to map
native format data from southbound interfaces into a standard format, that can be quickly
consumed by the correlation and analytics component.
COA platforms will deliver quick visibility to business problems, unlike traditional
operations tools, which solve distinct problem within a silo. In order to achieve this end
goal, we feel that the supporting taxonomy and canonical data model needs to be built
using business constructs instead of technology constructs to accommodate the addition
of new features in a scalable and cost effective way. It will be built on three primary
pillars – business, application, infrastructure – to address the operational needs of
three primary end user segments – business operations, IT operations and network
operations – via a common platform. These user groups that traditionally operated in silos
can now get the relevant insights using a common platform where targeted contextual
insights are delivered based on user profile and policies using role based access.
Business Process
Model Hierarchy
Example-1
Industry: Retail Banking
Business Process – Personal Loan
Example-2
Industry: Telecom
Business Process – Customer Care
Business
Forecast vs. Actual Map, SLA violation,
Transaction Troubleshooting, Root Casue Analysis
Forecast vs. Actual Map, SLA violation,
Transaction Troubleshooting, Root Casue Analysis
Forecast vs. Actual Map, SLA violation,
Transaction Troubleshooting, Root Casue Analysis
Forecast vs. Actual Map, SLA violation,
Transaction Troubleshooting, Root Casue Analysis
Collection of linked activities for personal loan
Forecast vs. Actual Map, SLA violation,
Transaction Troubleshooting, Root Casue Analysis
Receive Application, Check Credit,
Negotiate Loan, Choose loan
Add Service, Process Order, Submit trouble ticket,
View/Pay bills, increase QoS
Business
Applications
Loan Originating, Risk Management System,
Enterprise Content Managament
Billing, Ordering, Provisioning, Credit and Collection,
Product Catalog and Pricing, Suppy Chain
Application
Services
Tomcat, Oracle, SAP
Oracle CRM, Weblogic, Cassandra,
Apache Web Server
Business Process Analytics
and SLA Assurance
Business
Process SLA
Business Process
Definition
Application
Business
Services
Infrastructure
Infrastructure
Entities
Business Services
Business Services
Servers, VMs, Hypervisor, Switches, Routers, Firewalls, Load Balancer,
Storage Devices (physical and Virtual, On-Prem and hybrid).
S T R AT E G I C D I R E C T I O N S
Infrastructure:
The “infrastructure” category covers compute, storage and network components. This
includes physical as well as virtual infrastructure components such as servers, hypervisors,
VMs, switches, routers, firewalls and load balancers. It covers a wide range of vendors and
the associated syntax to communicate with each and convert the discovered data to a
common normalized format. Standard network collectors will implement algorithms to
discover metadata from the network devices, network flows and related paths.
Application:
Applications glue the infrastructure layer with the business layer of an enterprise. Critical
business services are hosted inside domain specific applications and are leveraged for
successful execution of key business processes. For example, a “Credit check” service may
be hosted in a “Credit and Collection” application and is used to fulfill multiple business
processes for different business units within an enterprise.
Business:
Each line of business (LOB) within an enterprise is measured against business SLAs, which
typically include metrics such as number of transactions (throughput) within a specific
timeframe (hour/days/months), response time (end-to-end) and success rate of the transaction. The LOB personnel pull ad-hoc business reports using a Business Intelligence (BI)
system to compare actual vs. expected business performance as it pertains to the transactions for their business unit. Sometimes a data center outage or a server outage or a
slow network can significantly impact the business and prohibit the line of business from
meeting their revenue target.
Having deep visibility into all of the infrastructure resources which are specifically associated with each business process will go a long way to proactively identifying business
impacting issues and in many cases completely prevent them. In addition it helps proactively plan for capacity needs, identify bottleneck or even troubleshoot a specific business
transaction for a customer.
Putting Cloud Operational Analytics in Context
COA platform will provide integration, correlation as well as analytics platform that
enriches the core values of APM, NPM, ITSM and SEIM systems to deliver value added
services for the business layer. Traditional APM or NPM systems don’t deliver northbound
functions that map operational data to business process components. For example – an
APM system detects that a business transaction is degraded due to a fault or performance
issue inside an application layer, but it doesn’t have the deeper understanding of business
process impact or deep visibility into infrastructure dependencies for the transaction.
Similarly, NPM systems focus on network performance data but lack business or even
application service context.
S T R AT E G I C D I R E C T I O N S
COA will deliver additional infrastructure, application and business level intelligence
to SEIM platforms with rich insight for security assurance and threat as well as anomaly
detection and management. Additionally, it enriches the value of traditional ITSM systems
by providing real-time insight into business process, applications and underlying infrastructure for root cause analysis while the ITSM systems focuses on orchestrating the
problem management activities such as ticketing, triage and problem resolution.
Business
Intelligence
(BI)
Cloud Operational
Analytics
(COA)
Lorem ipsom dolores deo
illiad magnetic deo illiad
deo exciteum deote
Network
Performance
Management
(NPM)
Application
Performance
Management
(APM)
IT Service
Management
(ITSM)
Security Event
Information
Management
(SEIM)
With all this in mind, OA is positioned to be the critical orchestrator across existing operations management domains. In many scenarios with traditional tools already in place
these systems will co-exist. Traditional tools play critical role in delivering the required
operational data with the OA platform delivering the front-end business centric visualization powered by strong correlation, real-time and predictive analytics. But mid-market
companies with more streamlined IT and network operations OA platforms may collapse
these separate management functions into a single integrated capability with a single
logical database combined with higher level reporting and analytics from each of the
prior domain areas. This ensures most efficient use of data, improved productivity and
addresses the operational agility for the enterprises. It may also deliver this with a lower
cost structure.
Fully converged or not, the COA platform will supercharge the NPM, APM, SEIM and ITSM
functions by leveraging relevant data and insights available within them and enrich their
value for the end business. The COA platform will be able to detect what, when and why
a critical business process is impacted due to an underlying issue either in the application,
network or storage layer. We call this concept “OI to BI” as it aims to correlate and connect
the operational intelligence systems to business intelligence systems for a 360-degree
view for business.
S T R AT E G I C D I R E C T I O N S
FixStream Approach:
FixStream Meridian platform aligns with this forward-looking OI to BI vision of Cloud
Operational Analytics. Though we focus on building our own data collector for the initial
release, our target is to enable data ingestion from many different sources including
existing systems within the ecosystem. We will establish an Open data ingestion
platform to normalize disparate data coming in from various data sources and systems
to a normalized format. Meridian Open data ingestion Platform will provide a framework
to interoperate with legacy vendor solutions for backward compatibility. As the data
traverses northbound within the Meridian platform, the processing shifts from syntax
based (specific to each vendor) to semantic-based (normalized) as the data processing
as well as persistence becomes standardized using Meridian’s canonical data model as
depicted in the following diagram. Meridian data model is highly extensible. It’s logical
definitions for base entities, metadata and relationships enables encapsulation from the
diversity and complexities that exist in the data collection layer for different vendors,
syntax, specifications.
Relationship
Base Entities
Topography
Business
Entities
• Entities are correlated by
relationships
• Metadata is collected data
where as relationships are
derived from metadata of
the entities
Application
Entities
Containers
Map
Performance
Fault
Transactions
Link
• Entities are characterized by
metadata
Metadata
Logos
Events
Infrastructure
Entities
Alerts
Ticketing
Our goal is to focus on building richer business focused correlation, analytics and visualization algorithms which is what end users most need, while commoditizing the data collection layer and working with the community to address the data integration complexities
for various compute, storage, network vendors as well as ecosystem components.
FixStream as a COA company aims to advocate and establish an “Open Data Ingestion
Platform” that will allow customers, network vendors, partners and systems integrators
to develop data collectors to interface with Meridian standard APIs for fault, performance,
transactions, incidents etc.
S T R AT E G I C D I R E C T I O N S
Visualization
(Dashboard, Integration Maps,
Personal, Interactive, Collaboration)
Data Analytics Platform
Extensible Data
Model
Indexing
Streaming
Analytics
Contextual
Correlation
Predictive
Analytics
Standard Meridian APIs
Device
Application
Performance
Faults
Events/Alerts
Tickets
APIs
Logs
Open Data Ingestion Platform
(Data aggregation, abstraction, transformation)
SSH/Telnet
Physical
Network Devices
Config
Files
Server, VMs,
Hypervisor
Packet
Capture
SNMP
Controllers
(Openstack, SDN)
VNFs
The company’s technology roadmap includes delivery of standard APIs and an SDK within
the Open Data Ingestion platform to enable quick integration with other domain specific
vendors providing meaningful data and insights. At a very high level, Meridian standard
data ingestion APIs will cover the following key logical domain functions.
•
•
•
•
NPM – Discover the network topology and network performance data
APM – Get application and business service level visibility to be able to tie it
to business process for business SLA management
ITSM – integrate to provide collaborative troubleshooting to further reduce
the MTTR, contextual and correlated troubleshooting flow
SEIM – Subscribe to real-time security alerts generated by SEIM vendors for
network hardware and applications
Consistent with the shift to hybrid cloud, Meridian will have APIs, which enable it to
detect and visualize applications running across multiple data center and control
scenarios with public cloud included.
S T R AT E G I C D I R E C T I O N PA P E R
To conclude, FixStream’s vision is to deliver a next generation Cloud Operational Analytics
platform that promotes Open Data Ingestion to enrich the capabilities of existing operations systems within the ecosystem. We are currently integrating with leading vendors
in this space. We will be a strong promoter of open ecosystem with a simple objective in
mind – make the business more successful and respond to the business needs proactively.
There are lots of data available, let’s take advantage and unleash the business value it
offers.
FixStream, Inc.
2680 North First Street
San Jose, CA, USA
www.FixStream.com
© 2015 FixStream Inc.
All Rights Reserved.
Contact Us:
FixStream is excited to be working with visionary technology
and channel partners and customers. We invite you to contact
us and experience the next level of visibility in your data center.
Email: [email protected]