Architecture

IBM Cloud applications use a proven topology based on DB2® and WebSphere® Application server.

Architecture details

IBM Cloud applications use topology based on DB2 and WebSphere. For more information on this topology, see: Recommended CLM deployment topologies 5.0

Note:

You don't have to pay for licenses that support IBM Cloud architecture such as Operating System, database, or application server.

Set up and hypervisor technology

IBM Cloud uses bare metal VMWare ESX and Quantistore servers.

Your work environment

Your work environment does not require any special system changes: any machine capable of supporting an Internet browser will suffice. Your system needs at least 4 GB of RAM and a minimum of Intel Pentium 4 processor speed.

Bandwidth requirements

IBM Cloud applications include a base of 100 Mbps of bandwidth with each SaaS as a base, with levels up to redundant 1 Gbps as an add-on cost. IBM® works with you to understand your projected usage and monitors throughput to identify bottlenecks.

Account management

You receive a dedicated Active Directory account and you can manage users by using Active Directory.

Patches and upgrade service

Product upgrades and fixes are part of the base service when new product versions are made generally available. The SaaS operations team runs extensive test scenarios to ensure a smooth upgrade before they conduct production upgrades. Upgrades are applied within 45 days of the equivalent on-premise product eGA. You are given seven days’ notice for major upgrades. For minor releases, the product is automatically upgraded.

Integration with other offerings

You can integrate your IBM Cloud applications with products like Rhapsody® or Rational® Reporting Engine that are installed in your workstation. If you want to have IBM host other products in the same environment as your IBM Cloud applications, you can use Customized Managed Services.

Back-up and business continuity service

IBM Cloud provides data backup by default. Incremental nightly backups are taken Monday to Friday with synthetic full backups taken on Saturday and Sunday. Each environment includes at least five recovery points with one recovery point daily. Additional restore points and additional backup targets (onsite or offsite) are available at additional cost. All backups are disk-based and stored in an alternate IBM Cloud data center. These backups are for Disaster Recovery (DR) purposes. The DR RPO and RTO are 24 hours and 168 hours respectively.

IBM Cloud has business continuity plans in place to provide for the recovery of both the cloud service, and the associated client content, within days in the event of a corresponding disaster.

Professional instances are backed up within the same data center using IBM Cloud eVault. Enterprise instances are backed up to a different data center.

Sometimes you need your data and its backup to remain in your country: it is standard in professional instances that the data remain in country, in the same data center as production data. In enterprise instances whether it remains in country depends on the data center chosen. For example, a USA data center is backed up to another data center in the USA. UK is backed up to a data center in France. It is possible through a statement of work for an Enterprise subscription to pay for setup and maintenance of a backup to remain within the same data center.

Subscription expiry

When your subscription expires or is terminated, you have a few options for dealing with your data:

Maintain your project data on an existing SoftLayer® infrastructure. You can work directly with IBM SoftLayer for Cloud hosting.
Receive a backup copy the latest version of the backup data for either archive purposes or on-premise installation.
Arrange data migration as an add-on service by using a separate services contract
Request that your data be destroyed in a secure fashion.

Supported data centers

IBM Cloud supports a subset of SoftLayer data centers. Data centers supported by these offerings are located in:

Washington DC
San Jose, California
London England
Frankfurt Germany
Melbourne, Australia
Softlayer Federal Data Center

Note: IBM deploys the Enterprise or Professional versions of IBM Cloud applications in a virtual private cloud (VPC) to data centers in the US, UK, Germany, and Australia. IBM will consider these deployment for you based on your specific requirements.

For more information about Softlayer data centers, see: Global data centers for your global business.

OSLC integration

OSLC support remains the same whether you run your applications on a desktop or in the cloud.

Integrating on-premises Jazz® deployments with IBM Cloud applications

You can initiate integration of your on-premises environment with your IBM Cloud applications, and you are responsible for establishing the integration between the environments. You may need to set up additional rules to open any firewalls to the environment.

Replication for higher performance (for users dispersed worldwide)

IBM Cloud applications are not designed with database replication capability for distributed sites. The closest approach to this would be separate instances (each in a different data center) with friend relationships for cross-project instance linking. As for performance, the system itself performs equally well in any data center. The issue is network latency and, most importantly, bandwidth. Although latency is certainly affected by distance, it is also affected by the number of network hops. Bandwidth is an issue on the IBM end; bandwith is always constrained by your enterprise network. Another option is to set up multiple site-to-site VPNs to get the traffic onto the SoftLayer backbone.

Additionally, you can use WAN accelerators or proxy cache server options. Both of those require deployment on your network, and you can request infrastructure support and guidance. All these options incur additional cost. Before IBM can explore any setup option, you must provide details of your user population locations, and you probably conduct tests and collect data to identify the most likely problem spots and the best solution.

Customizing your IBM Cloud environment

You have full administrative permissions, and thus you can manage the cross-tool integrations and simple data migration of your environment. You don't have developer access to the underlying code, application servers, or databases in the IBM Cloud applications. If you have extensive customizations that you want to bring forward into a IBM Cloud application, you may need Customized Managed Services, depending on the nature of the customizations required.

Testing IBM Cloud data center latency

You can test your IBM Cloud data center latency. For detailed instructions, see: Testing IBM Cloud Data Center Latency

Deployment

Rolling wave of deployment

You might want to deploy IBM Cloud application capabilities in phases, with a rolling wave of deployment. You want to become comfortable and adept with the deployment of a single capability before you roll out the next capability. With this rolling wave approach, you may be completing the full-scale deployment of one capability, while you are ramping up deployment of a second capability, and piloting a third capability. The key is to know, understand, and take ownership of each capability before moving on to the next capability.

IBM Cloud performance benchmarks

For information about performance benchmarks, see the 2012 IBM Cloud applications sizing guide: Collaborative Lifecycle Management 2012 Sizing Report (Standard Topology E1)

Scheduled maintenance plan

Patches and upgrades which are normally applied within a scheduled maintenance window. Sites have 24x7 network operations center monitoring. We have hardened the operating system and firewall, and we have NESSUS and anti -virus scanning to ensure that the environment remains protected against new threats. Operating system patches are applied weekly. If product security patch alerts occur, IBM notifies you and applies the patches as soon as you can accept the outage.

Safeguards for security in Cloud applications

IBM Cloud operations team members undergo background checks and are under confidentiality agreements.
Annually, the systems are tested for security vulnerabilities by a third party.
All SoftLayer data centers are highly protected with biometric access.
No one other than SoftLayer operations personnel, have access to the physical devices, except in the event of an outage if it is necessary to swap equipment.

Protecting and returning your data

Your data can be provided either on demand or in case of contract termination. In case of contract termination, we would provide you with the most recent backup of your data. If you want to receive data on demand, you may execute a data export of the environment to capture the data for import.

Data remanence after customer data is removed from storage

IBM Cloud infrastructure is based on the VMWare images. When you terminate your contract and transfer your data, IBM decommissions the images. The data is gone at that point. The data is destroyed in a secure fashion; prior to the decommission of the image, you will receive the most recent backup of your data.

Scalability of applications and data hosted in the IBM Cloud environment

IBM Cloud architecture is built to support both vertical and horizontal scaling. In other words, IBM Cloud has the ability to add additional hardware resources such as disk storage, memory, and CPU, and to extend an application with additional severs.

Incident and problem reporting

Any of your support calls are initially directed to IBM support, and they route the call based on the type of support needed. For example, if you have a product issue, your call is routed to the product support team. If your issue is about infrastructure, it is routed to the SaaS services team for further troubleshooting work, and they provide log data as needed.

Note: Even in the case of product issue, the SaaS services team is involved in the PMR management for professional- and enterprise- tier customers because a tremendous amount of work is needed to deal with log files and testing.

Automatic alerts for failure, degraded service, or exceeded planned utilization

The SaaS services team receives internal automatic alerts, and the team deals with these situations as they arise. Current monitoring covers the environment (CPU, RAM, disk space, user data storage), and the application to ensure access. When an alert is generated, the operations team begins immediately to address the issue. For CPU, RAM and system Disk space, you are scaled up as needed without additional cost. When you are close to exceeding your allocated user data storage, the SaaS services team adds 100 Gigabyte increments as a monthly cost for the environment. If you exceed your unique user subscription, you are billed for the additional users the following month. The sales representative is also notified of subscription overage, and will contact you to ensure that the subscription is sized to the right number of unique users.

Mitigation strategies for potential availability and performance issues

The SaaS service team monitors at the infrastructure and application level for any change in performance. If a change is detected, logs and resource levels are analyzed to determine the best possible remediation. Problems outside of the SaaS services team's control, such as internet down time or low bandwidth on the customer side, will be diagnosed. The SaaS services team can be very helpful in recommending temporary and long term solutions, but you have the responsibility to provide proper network and bandwidth for the service. Standard usage does not have significant bandwidth needs.

Redundancy features that ensure availability and performance

SoftLayer has redundancy built into its virtualization platform. If the physical hardware (SAN or bare metal server that the compute instance is on) goes down, it automatically fails over to another in the same datacenter. The redundancy applies to all tiers. For large and enterprise tiers, there is an additional layer through implementation of a virtual private cloud architecture that provides a level of redundancy and failover to deal with hardware issues. Underlying blades can be swapped out without impact to the service being provided. At this time there are no additional benefits to clustering that outweigh the risks and our existing customer base has seen outstanding reliability with the VPC approach.