Performance & Monitoring

Network Performance

IBM SRE Maximo and TRIRIGA instances are hosted in IBM Cloud Data Centers that deliver high bandwidth and low latency connections to our customers. SRE clients are currently hosted across thirteen (13) IBM Cloud Data Centers in the following locations: 

These locations are configured to IBM standards for provisioning Maximo and TRIRIGA SaaS. They also meet ISO certification standards and other security level requirements. 
 

IBM Cloud Network Points of Presence (PoPs)

Open

IBM Cloud Data Centers and network points of presence (PoP) have more than one 1 Gbps or 10-Gbps connections to top tier transit and peering network carriers. IBM has over 30 network PoPs worldwide. Network traffic from anywhere in the world will connect to the closest IBM Cloud Network PoP and travel directly across the IBM network to its primary data center location, minimizing the number of network hops and handoffs between providers. See the complete network map here. Inside the IBM Cloud data center, 1 Gbps or 10-Gbps network bandwidth is provided to individual servers via a pair of peered, aggregated front-end customer switches (FCS). These aggregated switches are attached to a pair of separate routers (that is, front-end customer routers, FCR) for L3 networking. This multi-tier design allows the network to scale across racks, rows, and pods within an IBM Cloud Data Center.

Network Bandwidth Requirements
A rough guideline for estimating network bandwidth requirements (Client Browser to Application Server) is 6.2 KBps (Kilobytes per second) per active concurrent user (Maximo) and 30 Kbps per active concurrent user (TRIRIGA)

Testing Network Latency
IBM provides low latency, high bandwidth connections over the internet to our data center locations using multiple network points of presence (PoPs). A looking glass site is available for customers that want to conduct informal ping, traceroute and BGP testing from IBM Cloud data center locations.

IBM Cloud Data Center Performance

• Best in class high performance server hardware

• SAN based storage

• Minimum of 40 Gbps of connectivity to each rack

• Redundant Public Network made up of multiple Tier 1 carriers for inbound / outbound traffic

• Multiple Network Points of Presence (PoPs) worldwide allow most direct path to your primary data center and IBM on Cloud environments, minimizing latency

• Separate Private and Management Networks for IBM internal server management run on separate, isolated NICs, increasing reliability and reducing public network traffic.

Application Server Performance

The IBM SRE team sizes and configures Maximo and TRIRIGA Application Servers for optimum performance using IBM established best practices. If your customer has an existing production environment, we use their named (licensed) and concurrent user information as well as their current server specifications to size an appropriate environment within IBM Cloud. IBM SaaS production environments leave sufficient headroom for unanticipated overhead. Production environment configuration includes appropriate server core CPU, Memory and disk sizing, JVM logical clustering and/or load balancing for User Interface, Inbound Integration, Reporting, Mobile and Batch workloads as well as appropriate JVM tuning. Dedicated virtual (or bare metal) machines are used to isolate and provide optimum throughput. It is the responsibility of the IBM SRE team to provision environments that will provide optimum stability and performance for the customer and maintain our availability SLA. This is based on the number of users, environments and components defined in the customer's IBM on Cloud contract.

Database Server Performance

The IBM SRE team sizes and configures IBM DB2 and Oracle database servers for optimum performance. This includes using IBM established best practices for initialization settings and indexes as well as empirical experience garnered from working with our existing customer base. Databases are maintained by our DBA staff to insure proper database health.

RUNSTATS are run automatically, REORGS are run manually when needed.

Index tuning is done when deemed appropriate in case of performance issues detected either via database monitoring, a SRE DBA or reported by the client.

Database parameters are configured for best performance as OLTP tuned for Maximo or TRIRIGA workload. Changes to parameters are applied when required due to workload specifics or as a result of performance alerts.

TRIRIGA Performance Summary

An IBM TRIRIGA Performance Summary on Linux, WebSphere and DB2 is available below. This is not IBM on Cloud specific, but is a close match to IBM SRE standard architecture:

Open

Monitoring

The IBM SRE team monitors of all IBM Maximo and TRIRIGA SaaS customer environments 24 x 7. This includes:

• System Availability (up or down) - HTTP Uptime - Servers are polled on a regular basis to insure uptime. Alerts are triggered to the IBM SRE team for any servers that are unresponsive or reporting HTTP 404 or 500 errors

• Storage - Disk and swap space is monitored on all servers including Total, Used and Free
  Alerts are triggered to CDS team if available space on any server goes below 25%

• Processor Utilization - CPU utilization is monitored on all servers Total CPU, User CPU, System CPU, IO CPU. Alerts are triggered to the IBM SRE team if utilization exceeds established thresholds over a 10 minute average

• Memory Usage (system cache) - Total, Used, Free

• Network - SoftLayer (IBM Cloud) for Network Monitor Alerts

• Firewall and VPN - Netorc for VPN alerts, LibreNMS for firewall resources (Memory Utilization, CPU, Interfaces Status (Up/Down)

• Thread / Process #

• WebSphere JVM availability, CPU Usage, Heap Usage, Web Container (Hung Threads, Active Threads), Web Module (requests per minute, avg respond time per request)

• WebSphere JDBC connection pools

• SaaS Performance Insight Dashboard (SPID) built on Grafana (used internally by IBM CDS)

• Maximo - Log In Users, Users per JVM (Maximo 7.6 and above only)

• TRIRIGA - Log In Users, Workflow Event Count

Other Monitoring:

• Backups - Backup status is monitored. Alerts are triggered to the IBM SRE team if automated backups do not complete successfully

• Synthetic Monitoring of application business case scenarios (on a per-request basis)

• Transaction Timing Reports (on a per-request basis)

• Activity Logging/Auditing is monitored for suspicious activity on IBM on Cloud systems using IBM's QRadar SIEM (Security Information and Event Management) system.