Untitled

• Open issues in cloud cover cloud computing course topics

Overview

• Cloud is not a universal solution for all IT service consumers or applications
• Cloud computing inherently involves issues not unique to the cloud itself
• Complex computing systems are vulnerable to hardware failures and security breaches
• Issues tied to cloud computing are discussed in relation to NIST USA guidelines

Computing Performance

• Real-time applications need high performance and predictability
• Cloud computing shares performance issues with other distributed computing forms
• Latency, measured by round-trip time, is unpredictable in internet-based communication
• Offline data synchronization on the cloud presents a problem, needing version control and group collaboration
• Legacy applications must be updated to leverage the scalable computing capacity of cloud
• Consumers need control over the data lifecycle and data intrusion/unauthorized access

Cloud Reliability

• It refers to how likely a system will provide failure-free service over a time
• Cloud reliability depends on the provider's infrastructure and subscribed service connectivity
• Cloud reliability measurement faces difficulties due to the intricacies of cloud procedures
• Internet unreliability affects security of cloud reliability

Cloud Reliability: Economic Goals and Evaluation

• Cloud provides economic savings upfront, eliminating maintenance costs and scale economies
• Cloud computing has economic risks
• Lack of automated SLA compliance mechanisms requires a common SLA template
• A common SLA template aids in making decisions about manual audits
• Portability of Workloads, initial cloud adoption barriers are the needs of data transfer to the Cloud
• Interoperability between Cloud Providers, consumers fear vendor lock-in due to a lack of interoperability,
• Disaster Recovery, economic and performance losses can be avoided for providers and consumer by using disaster recovery implementation

Compliance

• Compliance adherence is required to law and preferred procedure
• US Federal government agencies are creating methods to solve compliance issues
• Consumers are responsible for compliance, but providers implement it
• There is a limited visibility of security procedures, and consumers can ask about their implementation
• Acquire assurances from providers on data compliance and security issues
• Forensic support on any incident should be provided to evaluate attacks
• Providers are responsible for forensic analysis for SaaS
• Consumers are responsible for forensic analysis for IaaS

Information Security

• Info security is related to confidentiality and data integrity
• Info security is linked with the assurance of availability
• Administrative controls, technical controls, and physical controls can be used for security
• Application of administrative controls specifies the authority of specific users to create, update, delete, disclose, and transport data
• Technical controls cover Identity and Access Management (IAM), data encryption, and data audit-handling requirements according to any regulatory need
• Public and private clouds have security exposures
• Providers can provide physical data separation in addition to logical separation
• NIST recommends providers offer monitoring mechanisms for consumer security compliance

Disk Failure - Understanding the Threats

• Disk drives are electro-mechanical things which wear out and eventually fail
• Disasters or theft cause disk failures
• All mechanical devices have mean time between failure (MTBF)
• MTBF values calculated by manufacturers are generic across devices
• Traditional approaches have backup storage on storage, but risk data loss, and the recovery process takes a long time
• A Redundant Array of Independent Disks (RAID) has multiple copies of data and distributes them over disks
• Cloud-based data storage provides data access and replication over the internet
• Cloud-based backups are stored remotely and can be readily available

Power Failure or Disruption - Understanding the Threats

• Storms or power supply faults can damage computers with electrical surges
• Electrical surges can permanently damage data storage
• Unsaved data are lost during power blackouts
• Surge protectors are traditionally used, but useless in blackouts
• Data centers can install and use expensive uninterruptible power supply (UPS) devices or backup generators
• A solution for data centers is to shift data to another site
• Migrating data centers to Cloud reduces power backups as their costs are divided among consumers, and locations can be moved to another grid

Computer Virus - Understanding the Threats

• Surfing the web can cause the user to unknowingly download software or junk drives over their devices
• Computer viruses and spyware are risks to user's devices
• Install antivirus software and set to auto-update to get the most recent updates
• Restrict user privilege to install software
• Use a firewall over a router, on the computer, or around the LAN
• Cloud computing makes it difficult for non-cloud based viruses to penetrate
• Cloud providers have security for consumer safety of their data sofwat

Fire, Flood and Disgruntled EE's - Understanding the Threats

• Fire and fire suppressing practices can damage computing equipment, backup and data
• Unexpected rainfall can affect computing equipment
• Angry employees can launch viruses, delete files and leak passwords
• By residing the data center over cloud, prevents consumer effort of fire-prevention
• Floods cannot be avoided or put-off, avoid flood zones
• Cloud Identity as a Service can prevent terminated employees to do damage effectively

Lost Equipment and Desktop Failure- Understanding the Threats

• Losing equipment or laptops leads to data loss, or personal identity loss
• If confidential data is stored on the lost device, this can lead to increased security concerns
• Keeping backups, strong passwords, and login requirements is often used, and is effective
• Data can still be accessed via Cloud computing from an online face, or synced devices if lost
• User becomes offline until desktop in case of desktop failure is replaced
• Backup stores are traditionally used for desktops or enterprises and restore data as soon as possible

Networking Environment - Understanding the Threats

• Servers run the same risks as desktops
• Replacing blade servers is relatively simple
• Redundant Servers were once kept to replace the failed server quickly
• Cloud providers can provide 99.9% up-time through server redundancy and fail-over, and thus consumers aren't worried about server failure
• Network Failure can occur due to a faulty device and causes downtime
• Redundant networking and backup devices are important as a hotspot back up

Understanding the Threats: Database System and Phone System Failure

• Many companies store databases for customer records, sales, purchases, and HR systems
• Failed databases can cause dependent unavailable applications
• Companies either either use backup or replication of database instances
• Failover systems work with cloud-based database systems to minimize downtime
• If a phone system is damaged, employees can mobile phones, but customers are not
• Traditional solutions reduce Phone System impact
• Redundancy is internally provided via cloud based systems for reliable service

Measuring Business Impact, Disaster Recovery Plan Template

• Risks can be costly when reducing for resource redundancy and backups
• Risk-reduction mechanisms can limit investment
• IT staff should evaluate and classify each risk according to its impact
• Occurrence table to indicate disaster probability impact can be shown
• Document disaster recovery plan, to contain plan overview, goals and objectives, types of events covered, risk analysis and mitigation techniques
• Standard languages preferred for cloud services
• Plan created with clear resources consumer

Measuring Business Impact, Disaster Recovery Plan Template (Cont.)

• Cloud provider should act upon the disaster recovery plan of the consumer for critical data and software
• Compensation should be offered for interrupted service
• Implement controls to comply with legal, ISO, and audit standard
• Policies of provider should be sufficient to protect against malicious insiders
• Licensing, to be considered during licensing

Data Governance

• Ensure application interfaces are generic, before developing the Cloud based applications data adaptors
• Data Separation, ensure measures are implanted to ensure a seperation of sensitive and non-sensitive data from the client
• Data Integrity uses checksum and replication techniques
• Ensure providers follow data regulations
• Provide cloud consumers mechanisms to delete data and maintain evidence of proof
• Examine archiving procedures of the provider and ensure they are satisfactory

Security & Reliability

• Implement security policies along with implementing Consumer-side Vulnerabilities
• Encryption is applied for web sessions, data transfer and data storage
• Consumers should factor in security procedures of the provider for physical implementations
• If there is a physical attack on the provider's site, there must be a recovery plan
• Implement advanced authentication procedures by the consumer to prevent identity theft
• Identify access management, performance, and reliability scores, should the consumer use the key benchmarks after deployment

VMs, Software and Applications

• Sufficient mechanisms that make the provider implement to avoid attacks which implement the resources for VM's such as security and network should be implemented
• Plan for the VM migration, should the customers VM migration be planned across different providers if necessary.
• Cloud should not be used to implement critical software due to unreliable response time
• For secure development tools, integrate the security features
• Make sure the application has dependable function, and supports performance functionality
• Apps deploying in the could should be configured to be in VLAN segment
• The Cloud's security frameworks should have customizable and secured policies to the application deployed
• Standard programming languages should be preferred by consumers

Define System Goals And Requirements

• Migration to the cloud should be well planned considering
• Data security and privacy requirements
• A site capacity plan which allows initial operation for the app
• Scalability requirements at runtime
• System uptime requirements
• Business continuity and disaster requirements
• Operating system and programming language requirements
• Client devices support
• Reporting requirements, data export and API's should be specified

Protecting Existing Data and Know Application Charactaristics

• To restore at a certain time, it is highly recommended that the consumer back up data to help in restoring the data.
• Data privacy, storage and access, should there be any issues with discussion with the provider to be in the legal document
• The IT resources of what the application is should be required
• Important features must be lowed
• High and low demand periods
• Disk storage and RAM Requirements
• Usage requirement on bandwith, caching, database needs and simultaneous users

Deployment Schedule and Review Budget - IT Gov Issues

• Sufficient time for application after deployment should be given with a planned schedule
• Beta releases should be done to allow feedback given back to the developer
• Companies should license software, and use payroll key budget in relation to in house costs
• In-House maintenance, IT staff and data center staff may be factored in the final comparison
• IT Governance is important for defining the cloud solution with company strategy

Cloud based solution requirements for the customer

• Before deploying the application to the system
• System should consist functional and non-functional implementations
• Errors should be identified in any early stage, to save cost and time
• All requirements which are functional and non-functional should be specified by analyst

Designing cloud based requirements

• Designing solutions for logging and exceptions at critical parts for access
• Cloud solution should have accessibility through cloud, and use cases which are redundant
• Back up for downtime and restoring can be improved if data is handled through a re-negotiated provider

Cloud solution design metrics

• Existing and Future Capaciity to IT resources
• Interfacement of an os, device, and or browser
• Environment design and disaster mitigation consideration
• In terms of data design consideration Interoperabilty for Cloud solutions

Cloud Solution Design pt.2

• Performance - Using data replication, graphics and caching for use
• Price - Design with short and long term budget in mind with features of portability, recover, performance
• Privacy is to be implemented for medical, education and financial cases
• The Cloud should be portable with generic API's to work on Open Source applications
• Recovery is to be for disaster etc...

Cloud Solution Design Metric pt. 3

• Reliability - Design configuration for hardware failure issues
• Response time should be as fast as possible
• Robustness - Monitor cloud to ensure that little issues are tackled immediately
• Security is vital for a cloud solution
• Testability - the cases for functionality must have tests for functionality for the solution
• Usability - Ensure the customers can use the cloud for what they intend

Applications

• Cloud apps should scale to the maximum requirement
• The scaling process should be determined to meet new resources, and scaling on each resource that is on a down-time and up-time
• To ensure better scalability, balance design

Cloud scalability

• Optimize key pages, pages in the web should reduce space
• Selecting what part with code to measure is important as 1 to 20 percent of the code performs to to 80 percent
• Operations of the database with read and write can be designed to perform scaling
• The performance is not affected with read operations through repliacting databases
• Use databases for scalability operations
• What system has data log with performance should be considered with the log requirements of the user

Cloud Computing : Scalability And More

• Meeting the requirements, for scalable capacity is a must
• Scalability means acquiring addition resources to process the increasing workload.
• Don't scaling is to to be performed to increase the work load on each product
• Reducing performance, is tuning the process to a perfect solution of the RAM.

Scheduling and Resource Management

• Reducing the workload, execution, time and cost is vital for performance
• Resource scheduling and what it fulfills must be reliable, secure and scalable
• Cloud providers look to maximize profits while consumers look to time minimize execution

Scheduling resource Management pt.2

• Perform scheduling on budget with priority
• Ensure the clients and their requests are processed correctly
• Prevent the other constraint from taking priority

Time Scheduling of Recources

• Processing of time based scheduling makes for less stress on the request from a provider
• Schedule the recources with the job line approaching faster
• Evaluate through all stats such as number of deadlines, cost and overcost

Cost Based and Time Resource Scheduling

• Managing both cost and time may prove to be difficult if provisioning is used to meet the deadline on hand of the work
• Deadlines should be considered to have hybrid approach to maximize profits
• Cost Based scheduling should be applied with the Hybrid approach

Bargain Recourse Scheduling

• Consider resource market to see the what provider can offer the customer
• See what cost is there on the processing
• To meet all deadlines, bargaining is important and to succeed in bargaining

Profit base Recourcing scheduling

• Profit based scheduling aims at increasing the revenue.
• Must consider SLA requirements while following this process
• Also make sure the SLA is not broken

SLA and QES Resource Scheduling

• Focus on maintaining SLAs for scheduled recourses.
• Prevent overload to prevent overload
• All homogenous tasks should meet the SLAs from the start

Energy Rebaased Resource Scheduling

• Lowering the energy levels to decreasing costs and saving environment
• Estimations should be in sync with resource decision
• What distribution should be preferred with energy distribution for each task

Optimization Resource Scheduling

• Optimizing the algorithm to make it optimal for problems
• Must consider researcher inputs like bayesina and processing
• Drawbacks of time or the output may result some tasks

Priority-Based Resource Scheduling

• Lower high priority, in which which the resource to be allocated
• Lower priority tasks needs to be reallocatd to reduce starvation

VM Based resource scheduling

• The VM can host cloud based applications so it is deployed on a Vm Level
• Every application used on the VM can be accounted for scheduled requirements
• The Destination host may run out of IT required for previously deployed VMs