High Performance Cloud Native APIs Using Apache Geode

High Performance Cloud Native APIs
Using Apache Geode
Anna Jung Paul Vermeulen
1
@antheajung pvermeulen@pivotal.io

Unless otherwise indicated, these slides are © 2013-2016 Pivotal Software, Inc. and licensed under a Creative Commons Attribution-
NonCommercial license: http://creativecommons.org/licenses/by-nc/3.0/
What is Cloud Native….
2
● Cloud-native is about how applications are created and deployed, not where
● 12 factor manifesto outlines rules/guidelines for building a cloud native apps
● Must be a micro-service

Health Care Service Corporation
3
#6on Diversity MBA’s
50 Out Front for Diversity
Leadership Best Places to Work for Women
& Diverse Managers
Operating Blue Cross
and Blue Shield plans in
FIVE states: IL, MT,
NM, OK, TX
OUR PURPOSE
To do everything in our power to stand
with our members in sickness and in
health®
1936
year founded
+$1
billion
in IT
spend
Over
21,000
employees
15 million
members
2,100
IT employees
208.3million
claims processed
annually
LARGESTcustomer-
owned
health insurer in the U.S. and
4th largest overall

Business Problem
4
● Inquiry Service and Not Information Service
● Heavy weight, Hard to use SOAP services
● Constraints with Performance and Scalability

Architecture
5
Consumer
Consumer
Pivotal Cloud Foundry
Data Aware Functions
RESTful API
Pivotal Cloud Foundry
Micro-service is a collection of small services
- Implements business capabilities
- Runs in its own process space
- Communicates via HTTP API
- Deployed, upgrade, scaled and restarted independent of other services

Why Pivotal Cloud Foundry?
6
Cloud Foundry’s architectural structure includes components and
a high-enough level of interoperability to permit
● Fast application development and deployment
● Continuous integration
● DevOps-friendly workflows
● Integration with various cloud providers
● Integration with Spring frameworks for developer productivity

Why Apache Geode?
7
Apache Geode architecture provides low latency performance
and scalability at all levels
● Scalability of Geode cache
● Continuous availability of Geode
● High performance reads
● Caching patterns
● Supports cloning of micro-services for performance and scalability
● Provided isolation layer for making backing store changes
● Integration with Spring framework for developer productivity

Demo Overview
9
Client
● Act as independent client or Spring Boot
● REST API Controller
Server
● Client Function Provider
● Abstract Function Provider
● Function Test Setup
● Function Test
● Server-side Service Function
● Server-side Abstract Service Function

REST API Spring Boot Application
10
@SpringBootApplication
public class CustomerOrderClientApplication {
public static void main(String[] args) {
SpringApplication.run(CustomerOrderClientApplication.class, args);
}
}

REST API Controller
11
@RestController
public class CustomerOrderController {
@Autowired
private CustomerListProvider customerListProvider;
. . .
@GetMapping(value = "/customers")
public List<CustomerIO> customers() throws IOException {
return customerListProvider.execute();
}
}

Client Function Provider
12
public class CustomerListProvider
extends AbstractProvider<String, CustomerIO> {
private static final String FUNCTION_NAME = "CustomerListFunction";
private static final String REGION_NAME = "customer";
public CustomerListProvider(GemFireCache cache) { super(cache); }
public Set<?> getFilters(String request) { return emptySet(); }
public String getRegionName() { return REGION_NAME; }
public String getFunctionId() { return FUNCTION_NAME; }
}

Abstract Function Provider
13
public abstract class AbstractProvider<V, R> implements ServiceProvider<V, R> {
private GemFireCache cache;
private Provider provider;
AbstractProvider(GemFireCache cache, Provider provider) { . . . }
public Collection<R> execute() { . . . }
private Provider getProvider() { return this.provider; }
static class Provider {
Provider() { }
Execution getExecution(Region<?, ?> region) {
return FunctionService.onRegion(region);
}
}
}

Function Test Setup
14
@RunWith(SpringRunner.class)
public class CustomerListFunctionTest {
@MockBean private GemFireCache cache;
@MockBean(name = "customer")
private Region<CustomerKey, Customer> customerRegion;
@Mock private RegionFunctionContext regionFunctionContext;
@Mock private QueryService queryService;
@Mock private Query query;
@Mock private SelectResults<Entry<CustomerKey, Customer>> results;
@Autowired private CustomerListFunction customerListFunction;
@Before public void setUp() throws Exception {
when(cache.getQueryService()).thenReturn(queryService);
when(queryService.newQuery(any())).thenReturn(query);
when(query.execute(regionFunctionContext)).thenReturn(results);
}
. . .
}

Function Test
15
@Test
public void customerListFunction_returnsAllCustomers() {
ResultSender resultSender = mock(ResultSender.class);
when(regionFunctionContext.getResultSender()).thenReturn(resultSender);
when(results.asList()).thenReturn(getAllCustomers());
customerListFunction.processRequest(regionFunctionContext);
ArgumentCaptor<CustomerIO> captor =
ArgumentCaptor.forClass(CustomerIO.class);
verify(resultSender).lastResult(captor.capture());
CustomerIO customerIO = captor.getValue();
assertThat(customerIO.getId()).isEqualTo("customer1");
assertThat(customerIO.getName()).isEqualTo("Demo Person");
}

Server-Side Service Function
16
public class CustomerListFunction extends AbstractServiceFunction {
@Autowired private GemFireCache cache;
@Resource(name = "customer")
private Region<CustomerKey, Customer> customerRegion;
. . .
@Override
protected void processRequest(RegionFunctionContext regionFunctionContext) {
String queryString = "SELECT * FROM /customer.entries entry";
Query query = cache.getQueryService().newQuery(queryString);
SelectResults<Entry<CustomerKey, Customer>> results =
query.execute(regionFunctionContext);
//iterate over query result set
regionFunctionContext.getResultSender().sendResult(customerIO);
. . .
regionFunctionContext.getResultSender().lastResult(customerIO);
}
}

Server-Side Abstract Service Function
17
public abstract class AbstractServiceFunction implements Function, Declarable {
protected abstract void validateRequest(RegionFunctionContext var1);
protected abstract void processRequest(RegionFunctionContext var1);
public void execute(FunctionContext ctx) { . . . }
public String getId() { . . . }
public boolean hasResult() { . . . }
public boolean isHA() { . . . }
public boolean optimizeForWrite() { . . . }
}

Learn More. Stay Connected.
12/05 5:00pm Apache Geode Test Automation and Continuous Integration &
Deployment (CI-CD)
12/07 9:00am Main Stage Keynote by Mark Ardito
12/07 11:50am RDBMS and Apache Geode Data Movement
19
#springone@s1
p

High Performance Cloud Native APIs Using Apache Geode

More Related Content

What's hot (18)

Similar to High Performance Cloud Native APIs Using Apache Geode (20)

More from VMware Tanzu (20)

Recently uploaded (20)

High Performance Cloud Native APIs Using Apache Geode