Moving applications to a cloud is not for a gloomy of heart. Issues can stand adult that competence force we to re-architect a application, correspondence mandate can emanate roadblocks, and bandwidth problems can start if your cloud provider does not support low-level networking services such as multicasting.
After you’ve assessed that applications can run in a public cloud, there are other factors to cruise — configuration, information emigration and monitoring. What are some of a many common pattern tasks we need to keep in mind when migrating an app to a cloud? This checklist outlines pivotal points:
- Service catalogs competence enclose comparison versions of virtual machines (VMs) that are not amply patched for your needs. Use bottom images that have a latest confidence and organic patches.
- Review all daemons and services regulating on an picture and close down all images that aren’t needed. This will revoke a conflict use of all VMs.
- Configure firewalls between focus layers to minimize network trade between layers. For example, we should usually open ports that are indispensable to accept and respond to queries and support supervision protocols.
- If we need clever encryption, such as PCI DSS, safeguard your SSL/TLS encryption is strong. The customer and server in an SSL eventuality negotiate that naught apartment to use; they competence not use clever encryption if not configured to do so.
- When regulating a hybrid cloud model, configure a practical private network (VPN) to secure communications between inner and outmost resources.
- Use SSL server certificates to substantiate servers and support encrypted communications. Depending on your architecture, we competence be means to use domain turn or theme swap name (SAN) certificates to assistance minimize a series of certificates we have to manage.
- Use pattern collection such as Chef or Puppet to automate cloud supervision tasks.
Methods for relocating information to a cloud
In further to configuring cloud resources, you’ll need a devise to pierce information from a existent complement to a cloud-based system. There are several approaches to migrating data; a best choice will count on your enterprise’s mandate and architecture.
One proceed is to use database riposte to settle a duplicate of your database in a cloud before relocating a system. If you’re using a cloud for disaster recovery (DR), a reproduction competence already exist in a cloud. With this method, your prolongation complement stays live while changes are created to a replica. The time indispensable to switch from a stream prolongation complement to a cloud-based complement is minimized.
Bulk behind and revive is another proceed to information migration, that fits with non-transaction-based systems, such as information warehouses and other business comprehension (BI) databases used for information mining. However, this proceed can check a initial loading of a cloud-based database as good as incremental updates.
In addition, not all information is stored in relational databases. For file-based information stores, we competence wish to use rsync , a synchronization application mostly used for mirroring and backup. Rsync is accessible for Unix/Linux and Windows platforms.
Effectively monitoring cloud-based apps
Migrating applications to a cloud will extend your infrastructure and emanate additional resources we contingency guard and manage. To keep all applications regulating properly, we should:
- Include logs from cloud servers in a executive record supervision repository. You competence wish to say a apart record supervision repository in a cloud to revoke network trade and changes from a cloud to your inner centralized record supervision repository. Remember, though, this can criticise some advantages of carrying a singular indicate of entrance to record data.
- Use eventuality triggers to beget alerts about poignant events on cloud servers.
- Regularly use disadvantage scanning tools, such as Tenable’s Nessus, to check for confidence weaknesses on cloud-based VMs.
- If we are regulating a pay-as-you-go indication of cloud computing, automate a tracking of CPU use opposite servers and close down servers when use rates dump next a tangible threshold. Dynamically supplement servers when direct surpasses threshold rates.
- Use spot-pricing instances to save money. Amazon EC2 offers new ability during varying prices, depending on demand. For example, we make a bid to compensate a sold cost per hour for a VM, and as prolonged as your bid stays above a set price, we can continue to use a VM. This indication works good if we don’t need a pursuit to be finish on a set date or time. Spot pricing also requires that your pursuit redeem but losing too most work if a bid falls next a mark cost and your VM is close down.
Before migrating applications to a cloud, examination your complement architecture. Pay sold courtesy to network services and trade volume between inner servers and intensity cloud-based servers. When it creates clarity to pierce certain apps to a cloud, delicately cruise complement configuration, information emigration processes and ongoing monitoring and supervision procedure.
Dan Sullivan, M.Sc., is an author, systems designer and consultant with over 20 years of IT knowledge with engagements in modernized analytics, systems architecture, database design, craving confidence and business intelligence. He has worked in a extended operation of industries, including financial services, manufacturing, pharmaceuticals, program development, government, sell and education, among others. Dan has created extensively about topics trimming from information warehousing, cloud computing and modernized analytics to confidence management, collaboration, and content mining.
This was initial published in Dec 2011
Article source: http://www.pheedcontent.com/click.phdo?i=9e63928cab9a14e244aa267bc9c544e6