Cisco's SmartDB

 

Gold Stevie Award Winner 2016, Click to Enter The 2017 Stevie Awards for Sales and Customer service

 

Company: Cisco Systems, San Jose, CA
Company Description: Cisco is the worldwide leader in IT that helps companies seize the opportunities of tomorrow by proving that amazing things can happen when you connect the previously unconnected. At Cisco customers come first and an integral part of our DNA is creating long-lasting customer partnerships and working with them to identify their needs and provide solutions that support their success.
Nomination Category: New Product & Service Categories - Software
Nomination Sub Category: New Product or Service of the Year - Software - IoT Analytics Solution

Nomination Title: Cisco's SmartDB

Tell the story about this nominated product or service (up to 650 words). Describe its function, features, benefits, and performance to date:

As Cisco IT continues it’s journey towards the IoT digitization bandwagon, ability to predict events and implement corrective solutions on times has become the key focus. ‘Intensely Focus On Customers’, is one of the Cisco Values which we look to imbibe in our efforts to fulfill our perpetual dream of being the number one IT Company in the world. Without a robust and innovative mind-set, it’s impossible to fulfill the ever-growing customer demands.

Cisco is a giant IT enterprise which houses around 5400 production applications which run on top of 400+ production Oracle, SQL server & other cloud databases with 3200 hosts & utilizes 8PB of storage. On an average, 50 million rows of machine data is getting generated daily from this support infrastructure. Using this huge data set to achieve Operational Excellence seemed like a Herculean effort.

In spite of implementing various re-active and pro-active monitoring measures, our application customers still continued to experience service un-availability due to planned and unplanned downtime. Various enablers like automated alerting, self-service and self-healing helped reduce service un-availability. In FY15 the service availability was 98% and there was gap that had to be addressed.

More generally speaking, this was an industry opportunity for Cisco DB team to build a predictive analytics engine to perform predictive analytics and provide predictive alerts to enhance service availability to 99.99% and reduce the incident count by 20% year over year.

Our keen client focus and acumen for innovation drove us to come up with "IoT Analytics Solution - Smart DB ". The tool leverages predictive data models for effective capacity planning & management, future issue discover, Operation planning and improving data center quality.

In SmartDB, the data from various data sources like Oracle managed server repository, Unix hosts, storage frames, internal and external knowledge management systems are collected and aggregated based on defined Rules Engine.

The architecture comprise of multiple segments

1. Administrative Console
Provides a platform to configure and customize technology stacks that can be used across different layers

2. SmartDB Interface Modules
The interface module allows the tools to connect and consume data from different sources like Oracle Enterprise Manager (OEM), Social Media like Twitter and External Websites (via Web Crawlers)

3. Data Store
Repository for consolidation of data from multiple data sources from Interface modules

4. Smart Profile Manager
These are customized threshold per environment for accurate prediction. The smart profile has policies and rules engine
• Policies: Filtering criteria and aggregation
• Rules Engine: Algorithm to analyze the data based on customized rules

5. Service Adviser
It consist of the predictive orchestration engine which triggers the corrective actions based on alert and input data.
For example: SmartDB suggests us to combine these different infra activities to reduce planned downtime.
• Database - Security Patch downtime, DB upgrades and other maintenance activities can be combined.
• Hosting - Merge maintenance tasks and host patching and upgrades
• Network - Suggestions to overcome I/O bottleneck
• Storage – Storage consolidation activity which involves service downtime
• Application – Application load balancing, tuning suggestions

6. Data Visualization Layer
Intelligence Based Trending – Trending information for corrective actions

SmartDB – Use case implemented and Business Value :

Below are the details of use case implemented and the high level business value obtained

• Storage Prediction:
SmartDB predicted a reclaim of 92 TB archive storage space leading to a cost avoidance of $220K/Year.
86 unused servers identified and decommissioned, leading to a cost saving of $250K/Year.

• Hardware Capacity and Performance Prediction:
Hardware Capacity & Performance issues are being identified 2 Qtr. in advance and we fixed proactively (60% of Capacity & Performance issues are identified per Qtr.)

• Incident Prediction:
Incident of similar nature are proactively identified and addressed. 40% of critical incidents avoided last Qtr. Overall 15% of the incidents were reduced with different priorities.

• SmartBot
Smart predictive engine helps consolidate patches based on current industry trend which enabled us reduce planned downtime and enhance service availability to 99.99%

In bullet-list form, briefly summarize up to ten (10) of the chief features and benefits of the nominated product or service (up to 150 words).

With SmartDB we were able to achieve the following results:

a. Predicted unused space of 92 TB leading to a cost avoidance of $220K per Year.
b. 86 unused servers were identified and decommissioned. This helped us to save $250K per Year.
c. 60% of hardware capacity & performance issues are proactively identified and addressed.
d. 40% of business critical incidents were avoided in Q3FY16 using SmartDB predictions and overall 15% of incidents were reduced with different priority levels.
e. Reduced the planned downtime and enhanced the service availability to 99.99 from 98%
f. SmartDB was identified as one of the intellectual property (IP) use case for Cisco and we have submitted and won IP from ip.com