CASES

Case 1

One of the largest international
consulting companies

Platform operation

Challenge

The company had an acute task to change its processes — in particular, the process of developing services and applications. One of the key directions was to move away from monolithic architecture towards microservices.

Implementation of DevOps approach allowed to improve timing and save due to avoidance of proprietary software.

Solution

The Customer has improved inter- team and partner cooperation thanks to the new platform and flexible processes. New microservice architecture standard allowed to accelerate significantly bug fix and release of new versions.

The Customer plans to improve the current installation to develop and run new services on the Skipper Platform.

Solution

SMART technologies SOFT team, working together with the Customer, conducted a pilot project to deploy the Skipper Platform and migrate one productive application followed by load testing.

Next step was assistance to the Customer in migration of Open Source applications of interest to the Skipper Platform.

Case 2

Housing and Utilities Company

Challenge

Housing and Utilities Company actively implements IOT technology for smart meter management, monitoring and interaction with resource providers.

The Customer’s task was to make IT infrastructure management flexible and elastic with the increase of system clients and number of services.

Solution

The Customer received saving in infrastructure, since the development and testing of services is handled in the Yandex.Cloud on the Skipper Platform.

The Customer received an Enterprise-class Platform supported by vendor (24x7).

Solution

SMART technologies SOFT team created a workspace in the Yandex.Cloud, where the Сustomer performed functional testing
of one of its services.

The next step was to deploy the Skipper Platform on-prem in the contour of the Customer's data center and migrate analytics services, databases and other Customer's applications.

(maintenance of houses, schools, kindergartens)

Application catalogue on the Skipper Platform and Open Source tools supported by SMART technologies SOFT team reduced the risk of unplanned system downtime.

Case 3

The Skipper Platform
in a Large Retailer

Challenge

The Customer uses SAP HANA for hot storage and is starting to use Hadoop.

The average processing time in SAP HANA is up to a minute, the average processing time in Hadoop is up to 10 minutes.

SAP HANA Alternative Solution

The task was to propose an alternative solution with similar or better performance reducing costs and replacing import.

Solution

SMART technologies SOFT team handled Hadoop and SAP HANA testing in order to find out the acceptable speed range as well as collecting statistics on query duration and server resource consumption (2 parallel threads executing requests with different conditions). The aim was to propose the best Skipper Platform configuration meeting the criteria of performance and functionality.

Basing on analysis, SMART technologies SOFT team proposed a scenario to differentiate data access for users, allowing access to sales results only in their regions.

Solution

The results on the Skipper Platform are close to the results on SAP HANA, but the total cost of ownership is significantly lower, which allowed the Customer to get an import substitution and improve performance at the same time.

Timing:

Insert new and modified records: 17m records in 0:50;

Physically delete old versions of modified and deleted records in 0:24;

“Disability” of changed or deleted records: 17m records in 04:50.

Case 4

The Skipper Platform in a Telecom Company

Challenge

The Customer has set the task to calculate the margin from subscribers.

Solution

SMART technologies SOFT team proposed to create the margin calculation algorithm on ClickHouse Database (on- premise) using all advantages of the Cloud Platform.

Solution

Total timing of algorithm performance on the Skipper Platform: 25min 47s

Total run time on the Skipper Platform:

25min 47s

* In one of the steps, the consumption of resident memory by the ClickHouse server process was 90 GB

Current run time:

2h 25min 34s

Timing of Timetables Filling

For the largest branch with:

50 mln subscribers

300 mln in ACH

40 mln clients

2400 mln in ACL

0.3 mln in CORR