Introduction

Founded in 1938, Pemex (Petroleos Mexicanos) is the largest petroleum and petrochemical company in Mexico, with a workforce of approximately 150,000. Its core businesses include exploration, exploitation, refining, and sale of oil products, petrochemicals, as well as supply and sales of liquefied gas, natural gas, and basic petrochemical products. In 2012, Pemex posted a revenue of 1.0647 trillion pesos (about CNY532.3 billion; US$85.5 billion).

Challenges

In recent years, rapidly developing IT technologies have played an increasing role in energy companies, helping them improve resource utilization, safeguard oil operations, and reduce production costs. As a modern large energy company, Pemex has high IT requirements. However, Pemex's original data center was aging and inefficient. To support its business growth, Pemex urgently needed to migrate IT infrastructure from the live network to a new data center (referred to as the container data center in this project).

A survey of Pemex's original data center showed that 319 servers, 9 storage devices, and 63 network devices would need to be migrated. If all these devices were shut down simultaneously during migration, services could be interrupted for as long as 15 days before all procedures were complete, which would have a disastrous effect on Pemex's business operations. 

Pemex's manager in charge of the migration project said: "We hope to smoothly migrate the data center in the shortest time possible and with minimal service interruptions."

• Minimal service impact

Service impact was one of Pemex's primary concerns in the data center migration project. If multiple key applications were migrated by simultaneously shutting down related devices, Pemex would suffer from serious economic losses. Pemex required migration in a series of waves to minimize service impact, where each wave fully considered many factors, including service loads and inter-service correlations.

• High stability

High stability is a key requirement for IT systems in energy companies like Pemex. Many key applications, including a petroleum exploration and generation system and a petroleum pipeline surveillance system, must provide 24/7 uninterruptable services. As such, stability must be assured during data center migration. To do so, Pemex required a comprehensive migration plan containing detailed migration methodologies and contingency measures against exceptions, such as device failures or network outages.

Huawei Solution

• Migration in a series of waves, minimizing service impact

Huawei provided comprehensive data center consolidation and migration methodologies: 1) conducted a survey of Pemex's live network, 2) analyzed equipment features as well as service loads and inter-service correlations, 3) produced a service migration report and a migration feasibility report, 4) determined optimal application grouping choices, and 5) carried out migration in a series of waves.
In addition, different services were migrated using different policies; for example,

• All non-critical applications were migrated all at once during a single wave, and then critical applications were migrated in a series of different waves;
• Closely related services were migrated all at once;
• Services with a heavy workload were migrated in a series of waves. During each wave, only the minimum number of servers were shut down. 

Huawei also used a large Layer 2 approach to stretch the network, ensuring servers would still be reachable using original IP addresses after migration. As a result, service migration was truly transparent to the customer.

• Well-designed data center migration services best suited to the live network

After a detailed survey of the live network, Huawei classified equipment into different types and used different migration approaches accordingly; for example,

• Storage devices store large amounts of key production data. If such devices were shut down for a long period of time during migration, Pemex would incur serious losses. Therefore, the migration services provided by the storage product vendors were used because they would minimize the impact on ongoing services.
• HP and Cisco devices running on the live network carried small amounts of data. Shutting them down had little impact on ongoing services. In these cases, Huawei provided their own migration services. During migration of HP servers, Huawei servers were used as backup devices. In addition, the original data center and the new container data center merged to form a large Layer 2 network using optical fibers; Huawei switches and Cisco switches were interconnected at Layer 2 to ensure functional services at both the original and new data centers during migration.
• To reduce migration costs, devices eligible for free migration services within their warranty period were migrated by device vendors, while migration of all other devices was handled by Huawei.

• Huawei's comprehensive product offerings and superb planning capabilities, ensuring fast device replacement in the event of device failures

Huawei has a wide range of servers, storage devices, and network devices, capable of fully adapting to Pemex's live network. If devices failed during migration, Huawei could promptly replace them with new ones. In the event of an emergency, Huawei had contingency plans in place and related products to ensure a successful migration.