Hitachi has delivered a software, which is Hitachi Infrastructure Analytics Advisor (HIAA). HIAA delivers visualization, intelligence and automation to optimize infrastructure health while quickly identifying and troubleshooting performance issues.
In my first post, I introduced a use case to determine and solve performance bottleneck in Hitachi Unified Computing Platform CI (UCP CI) environment.
Hitachi Infrastructure Analytics Advisor – Hitachi Data Center Analytics SaaS edition (HIAA – HDCA SaaS edition)
Now HIAA – HDCA SaaS edition is available to you. This is a new consumption model of HIAA. Hitachi provides a HIAA instance running on the Public Cloud. But HIAA – HDCA SaaS edition provides same user experience as well as on-premises version of HIAA.
HIAA – HDCA SaaS edition relieves installation process. Hitachi provides the configured HIAA server instance on Public Cloud. Customer can start managing the system using HIAA GUI instantly when user obtain access information. Customers do not have to prepare HIAA server.
- No Physical/vm-based server required
- No installation and configuration
2. Maintenance Free
HIAA – HDCA SaaS edition is guaranteed its availability. Hitachi observes HIAA instance and keep it alive. Thanks to this service, customers can focus on managing their system.
- Instance Management
- Version Management
3. Subscription model (SaaS)
In HIAA – HDCA SaaS edition, subscription price is determined by the amount of performance information customer wants to upload. Thanks to this pricing model, customers can start using this solution in lower cost than the full upfront payment model.
Combination of HIAA – HDCA SaaS edition & UCP CI
HIAA – HDCA SaaS edition is better solution for UCP customers those who prefer low-touch infrastructure management. Because customers do not need to care HIAA – HDCA SaaS edition instance. Hitachi provides management service of HIAA – HDCA SaaS edition instance.
HIAA – HDCA SaaS edition provides server-less model to customers. This means customers can implement more higher density of user application and VMs.
This is the system overview. Basic components, which are Probe server and HIAA Server, are same as on-premises version of HIAA. But HIAA server move to Public Cloud. Generally, most of customers protect their system by the firewall.
1. Network configuration
a. The Probe server communicate with the HIAA server via HTTPS protocols. (No other protocols are required.) Please negotiate your network administrators to allow outbound connection HTTPS port 443. For example, changing Firewall setting or though HTTPS proxy.
b. In Public Cloud side, Web Application Firewall (WAF) allow to connect from user network. Hitachi have to list the public IP address of customer network. Please contact Hitachi representative.
2. Access information from Hitachi
Hitachi provide HIAA – HDCA SaaS edition access information to customers.
- HIAA Server information to upload performance information
- HIAA GUI access information to login console
I am going to introduce an example configuration, which is HIAA – HDCA SaaS edition manages UCP CI in the on-premises data center.
1. Configuration in on-premises data center
There are same processes to install and configure in the on-premises side. (Reference: https://community.hds.com/community/products-and-solutions/vmware/blog/2017/12/29/part2-end-to-end-infrastructure-performance-analysis-of-hitachi-unified-compute-platform-ci-simplified-with-hitachi-infrastructure-analytics-advisor-hiaa)
The only difference point is to switch data upload server into HDCA server which is a part of components in HIAA – HDCA SaaS edition. The probe server in on-premises have to connect HDCA. Once customer login to probe server, then input HDCA server information.
After a few hours, configuration and performance data will be appeared in GUI screen.
2. HIAA – HDCA SaaS edition configuration
As I mentioned before, customers can start to use HIAA – HDCA SaaS edition immediately. Here, I recommend making a “Consumer Group”. This is a nice feature to identify pods later.
1. Generally, administrators use vCenter to observe system health. If unexpected workload raise, administrator will notice that noisy neighbor issue happed. In this situation, administrator will check system performance in vCenter performance monitor. For example, CPU, Data Store and so on.
2. On HIAA GUI dashboard, there are an alert on the storage system.
3. Breakdown of the storage system, there are alerts on Cache and Parity Groups. Administrator have to find a root cause of higher CPR rate.
4. At this moment, HIAA provides analytics feature of bottle neck. BasePoint is the origin point that Administrators start to investigate. Now we are going to start from Cache because we need to check why CWP is so high.
5. HIAA shows suspects of higher workload. What volume has much high workload.
6. One more approach, HIAA can show how busy Parity Groups works. In this use case, to keep up user’s workload, Administrator decides to install more HDDs to enhance drive performance. Because Parity Group Utilization shows exceeds 80%. This shows Parity Group 04-03 and 04-04 is busy at this time.
7. What HDP pool should be added more HDDs? HIAA can show where PG should be installed.
8. After installing HDDs, Auto Rebalance will start working to distribute workload evenly. For a while, Auto Rebalance will continue working. After that, the system performance will be boosted. User can experience enhanced performance.
In this article, I introduced HIAA – HDCA SaaS edition. HIAA – HDCA SaaS edition is working on the public cloud. Hiachi prepare it for customers then the customer can start to use instantly. This provides lower OPEX and CAPEX of performance management to customers. Meanwhile, HIAA – HDCA SaaS edition provides same user experience as well as on-premises model.