(1) unclear leadership
The most common mistake in planning is failure to establish clear leadership.
This means identifying who is responsible for clear communication and leading the team during each phase of the migration process.
In a single department, the default is to adopt a leadership style that represents their best interests.
The project leader must be fair and impartial, and he needs to understand and accurately reflect the goals and standards of success for each team.
This person must also have good executive power and communication skills to keep everyone committed to the same goal.
(2) lack of a complete infrastructure assessment
The most common mistake in the exploration process is the lack of a complete infrastructure evaluation.
In particular, the documentation for each rack and each device is associated with the application.
This assessment should take care of all things, including physical and virtual appliances, network devices, network topologies, etc. Don't take shortcuts for not having these devices, and migrate too many information migrations.
Tip: include evaluating business and technology dependencies.
For example, applications, networking, and database server related applications must be migrated as a package.
(3) underestimate the migration time
A common mistake in actual migration is to not set realistic time expectations.
Production migration is inherently slower than testing migration because they require more care and attention to detail.
Tip: use test migration to evaluate the actual migration time.
You'll be surprised how long some applications will take.
Don't expect the migration to wait.
(4) setting, but forgetting
The most common mistake in the new infrastructure management process is the "setting (actually) forgetting" mentality.
Everyone was excited and wanted to move immediately.
Make sure that the technology and business go on steadily, without leaving out any details.
Tip: after completing the migration, take at least some time for active monitoring and support within 48 hours.
(5) think that the work is done
Phase one of the most common mistakes in size, is that now you're done, and lost momentum, these words may be in the process of migration, beyond the enterprise set up the annual plan, quarterly review, and AD hoc infrastructure requirements planning.
Companies have just invested a lot of time, energy, and money into a difficult process.
Now don't lose motivation or lose focus on detail because the migration is over.
This is not the only mode of migration for data centers, and these are certainly not the only ones to do it.
This is not the only mode of data center migration, but these are not all mistakes people make, and there are other mistakes.
The most important thing here is that in the continuous updating process, the enterprise's technical, operational requirements, and experience are all changing.
The preset data center is migrated to the IDC data center with several key considerations:
(1) location (location of physical location and IT personnel)
Like in a real estate market, location is always the most important.
For IDC providers, location means resources available in the region, and location has a significant impact on the security and good availability of data center assets.
The availability of local climate (flooding, extreme temperature fluctuations, storm frequency and intensity), earthquake history and critical infrastructure (such as lines, roads and airports) should be taken into account.
For industries that must be more stringent in compliance with regulations, such as the financial sector, it may be prohibited to provide customers with the business of storing data in international and even national border areas.
The same principle applies to supporting employees.
Whether or not you keep your staff or not, you need to know IDC's staffing requirements for IDC providers.
Although some IDC tenants can still keep the IT staff, other tenants, as part of the outsourcing contract, IT completely before the next step, according to oneself circumstance to establish a test procedure of scorecard, the purpose is to minimize potential suppliers, for the lease IDC space for customers is a pertinent advice.
(2) operation of cooling system
The cooling system in IDC space is essential after securing a fixed power source.
Power use efficiency (PUE) is critical in optimizing cooling cost and efficiency.
PUE can show how much it costs to deliver electricity to the rack.
Ideally, the tenant needs to multiply the power (measured) by the PUE coefficient to show the extra power required for cooling.
Look for a hybrid cooling technology (for example, using natural cooling) IDC to ensure the redundancy of the refrigeration system.
(3) have DCIM software
While DCIM software has become the standard for managing data centers, not all IDC providers provide complete technical support.
Because data centers have traditionally had a number of dedicated devices, these devices have a lot of complex technologies that might be problematic to manage.
Typically, devices have management software, but individual software systems may not be compatible or integrated.
This will lead to confusion in the data center.
So make sure the IDC provider has a DCIM software.
Do you know if all the systems in the data center are connected?
Are all sensors connected to software and monitored by these software?
Can they dynamically generate the dashboard and report and quickly move to the site, cabinet, and rack level?
Do they have an end-to-end asset management capability?
Have they been integrated into other ITSM systems (IT service management systems) and enable you to take advantage of the most needed functionality?
(4) physical security status
Without tight financial security guarantees, any bank would not be delusional.
The same is true of data center assets, which may contain only the most valuable assets other than human capital.
For truly secure facilities, insist on an internal security team.
On the analysis of the security model, be sure to assess internal security personnel, layered security zone, cameras, and security systems, these systems can be 360 degree coverage, and has advanced security certification, such as PCI DSS 2.0, ssae-16 16 and ISO 27002.
(5) power supply situation
Electricity is a broad problem and a narrow one.
On the macro level, the stability and redundancy of regional power grid infrastructure need to be considered.
Look for power stations, substations, cables to data center distances, locations and considerations for the entire distribution system.
Make sure there is no power limit to the IDC area.
This may sound far-off, but if you've ever experienced a power cut, or a power outage during peak or seasonal periods, you wouldn't think so.
Don't forget to learn about local power outages and maintenance records in order to make contingency plans.
At the micro level, it is necessary to consider electric power monitoring in IDC space.
Can they accurately quantify the power of measurement, whether it has the flexibility to increase or decrease the power consumption over time, thus making clear the current use of electricity?
Do they have a way to detect, monitor, and mitigate power surges and other anomalies?
What are their backup and disaster recovery plans when power outages occur in managed facilities unrelated to external power?
All this requires careful consideration.
(6) service level agreement (SLA)
By defining slas provide visibility and management tools for the tenant, this is the cornerstone of good continuous relations, can the initiative to avoid conflict, choosing the right hosting providers, create or have good SLA and to establish a clear demarcation line is crucial, usually, the SLA can according to the specific tenant demand and are managed to adjust the asset.
This means identifying key workloads, applications, servers, etc.
It emphasizes that when you pay for data center IDC, you buy critical infrastructure and ongoing maintenance, and you want to ensure that the SLA includes maintenance and testing.
Look for MOPs (procedural laws) and SOPs (standard operating procedures) with documented documentation, constantly using and improving.
Also make sure that they include the good management tools to monitor the power supply, cooling, housing situation, with the normal operation of the conventional time and status report of the environment, and provide the log aggregation tool, used to collect all kinds of server, system and security logs for analysis.
In this way, valuable time can be spent identifying and solving potential problems rather than buck-passing.
(7) workload and workflow management
After checked all the physical factors, now should focus on the work load transfer method and the working process of management, organization based on the data center attempt to provide the data or the type of application, there are several key consideration, cloud and big data will continue to develop and change the organization how to share data, especially when the distribution of information between multiple locations, you can now more than ever, and transmit information quickly and effectively IT vision is by "data on demand," BYOD (bring equipment) and the Internet of things (iot), and other important change shape, so you want to make sure that your hosting service provider in - but not only, also can keep doing this, so IT won't make your data center management outdated faster than the application.
Similarly, balance the workload, continuity and disaster recovery is vital for sustainability, data must move from amount of bandwidth and hosting providers offer may mean a great user experience and failure managed deployment, the difference between their workflow management system, or how good they are, can help to determine some of the data and infrastructure components delivery priority.
In addition, it will help to determine which need high uptime requirements, compared with low priority application in bottlenecks or emergency, you will be able to access the most important information in the first place.
The preset data center is migrated to IDC data center risk and countermeasures
Risk 1: service availability
The main purpose of the data center is to host applications that run services for the business.
Whenever you consider moving from one data center to another, you must first consider the availability of the underlying services.
These services include infrastructure applications such as activity catalogs and customer-facing applications such as SAP.
When a service from a data center to another data center, you must develop a strategy, when a specific service migration and application dependencies must be taken into account for each other.
One common way to ensure service availability is to create a migration group and then share the interdependent applications in the same group.
As for services that support most enterprise applications, such as active directories and DNS, a common practice is to extend these core services across data centers.
The service is still on both sides of the data center until the migration is complete.
Risk 2: data migration
The application data migration from one place to another may be a data center migration project one of the most complex link, a simple solution is to perform based on magnetic tape or disk backup, and perform recovery, however, a similar translation migration, backup and restore the ability to provide timely recovery services limited.
In addition, backup and recovery are not ideal for data migration best - it's better suited to a disaster recovery scenario with limited data recovery scenarios.
The main method of selecting for most data migration is to configure a leased line.
If there is a dedicated connection between the two data centers, the migration team can make full use of the hardware or software synchronization mechanism to perform data migration.
In addition to being able to migrate data, this method can be used to perform P2P migration, P2V migration, and virtual machines to virtual machine (V2V) migration.
Many companies decide to have multiple connections between two data centers.
Connection requires at least two lines: the connection to support all the way ordinary end users and data centers to the flow of data center, in order to support applications such as active directory and the application to the application of flow;
The second route, usually faster, is used to perform data synchronization.
A two-way connection prevents two completely different flows from interfering or affecting each other.
Risk 3: hardware migration
Migration physical servers usually have two strategies: one called "translation" (lift and shift), another is called data replication, in translation strategy, the hardware on the activities of the truck, and then install the new data center.
The system is backed up before moving to the new site, but this strategy carries some risks.
One of the biggest risks is that physical servers may be damaged during handling, and damage in the handling process can result in a useless backup.
Another challenge is that the two data centers are too far apart, so this approach is unrealistic and there is no guarantee that the service will be available within acceptable timeframes.
The second strategy is to move data through a leased line.
The leased line brings two sub-swing hardware solutions.
One solution is to perform physical machines to the physical machine (P2P) migration.
P2P migration requires the purchase of similar hardware so that the application and hardware of the original data center can be migrated to the past, while ensuring the shortest downtime.
The other hardware migration scheme is the physical machine to the virtual machine (P2V) transformation.
The P2V needs to convert the physical machine to a virtual machine by renting a line.
P2V aims to achieve two objectives:
The first goal is to move the workload from one data center to another, while ensuring the lowest cost of hardware.
The second goal is to transform the data center by moving to a virtual platform.
The P2V migration is a popular option because many engineers are already used to performing this transformation as part of the previous data center project.
Conclusion:
Will be migrated to IDC, the data center may be crucial for the business, look before you leap, must consider the main factors listed above, matters needing attention and risk, remember, all of the hosting provider is not the same, the more you can choose the best hosting providers, the more you get more value from the migration and more peace of mind.
沒有留言:
張貼留言