M630 NIC Options

4x1Gb – Broadcom & intel (new for 13g)

2 x 10Gb  - qlogic 57810, intel x520, and Emulex (same as available on m620)

4x10Gb  - qlogic 57840 (same as on m620)

In Q1cy15 we add new intel “Fortville” x710 controllers

2x10Gb

4x10Gb

DELL Storage SC4020 800GB Tier 1 SSD

We are pleased to announce the availability of a new drive – 800GB Tier 1 (Mixed Use) SSD – the first of its kind in an SC system. This drive type is to be used as a Tier 1 SSD similar to the existing 200GB/400GB write intensive (WI) drives. The SCOS will identify this drive with the same WI classification and will use the same tier as the 200GB and 400GB WI SSDs.  The industry is referring to these drives types as “mixed use (MU)” drives but from a Dell Storage perspective, these are used and tiered the same way as the Write Intensive (WI) SSDs.

Dell Storage is shifting to mixed use drives for a number of reasons:

1.       As new generations of SSDs are released, WI and MU drives will offer similar write performance.

2.      As capacity grows, MU drives offer similar endurance as the smaller WI drives when comparing total petabytes written in the drive’s life. 

3.      Field and customer data has helped determine that MU drives offer sufficient write endurance for even the most write intensive environments. 

4.      Mixed use drives offer higher capacity at a lower $/GB than comparable WI drives.

5.      The broader SSD market is making a shift to MU drives.

Table 1: Comparison of WI/MU/RI Drives for Dell Compellent

Dell Storage Use	Write Intensive		Read Intensive
Market Terminology	Write Intensive (WI)	Mixed Use (MU)	Read Intensive (RI)
Workload	Mainstream Applications Any usage		Mostly Read 90/10 R/W Mix
Used with Compellent	Yes		Yes
Capacities	200/400 GB	800GB	1.6 TB
Endurance (Full writes / Day)*	10-30		<3
Endurance (written PBs)*	Up to 30PB		8PB
Random Read IOPS*	Up to 20K +		14K +
Random Write IOPS*	11K +	8K +	4K +
Sustained Write Bandwidth*	200-250 MB/s	150-225 MB/s	50-100 MB/s
List $/GB	Up to $31	$16.60	$5.25

* These performance values are for individual drives during benchmark testing. These values do not reflect actual system performance values. Values are expected to differ once drives are managed in the system with RAID virtualization and other system functions.

It is important to note that we recently moved to a new warranty policy that protects SSDs in Compellent Systems for the full length of a system’s warranty, regardless of wear or maximum rated life.

Creating group of VLANs and assign ports to it

Creating Vlans should not be a problem. Try this:

interface group vlan 120-125, vlan 130-135

Once you created the VLANs, you can use the following command to tag a port to all VLANs you want:

int range vlan 1-4000

tag <port#>

Emulex OCm14102-u2-d in M620 and vSphere 5.5 Issue

If you use the latest 5.5u2 DELL iso from our ftp, boot up the host with either NPAR on or off, the nic’s in iDRAC will show up/up, the switch will say up/up, however vSphere will say Down, half-duplex.

To resolve this issue run esxcfg-vnics –a <vnic> to set it to auto-negotiation (the driver seems to have overridden this). This will make the nic’s come online. It seems to survive reboots.

Biggest challenges with most of the service providers are facing today

Biggest challenges with most of the service providers are facing today :
- No way / poor feedback capturing
- Lack in Product Management experience to improve based on the 3 party feedback (Internal, Customer and Partners/investors/vendors)
- Competing with Big fishes/crocs i.e. Azure and Amazon against their strongest game i.e. IAAS and PAAS and still trying to be competitive
- Not realizing which market to target or which one to avoid
- Considering Technology as the game changer for their offerings
- Forgetting the importance of Enterprise Architecture / SOA within this environment which will become the important part to innovate later
- Poor analytics on the market requirements, their marketing campaigns, long term market shift
- Doing large CAPEX instead of considering Partner model (use companies i.e. ServiceNow etc which can provide SAAS based Service Management Solution instead of investing money from their pocket)
- Lacking domain expertise and lacking domain and country regulations
- Lacking architecture experience in such a massive and complex environments
- Resources capabilities and Roadmap
- Running operations in a similar fashion the way they run in single organization. Lacking understanding of ISO 27000:7
- Forgetting International Standards and importance of them to bring competitiveness : ISO 27001/2 , ISO 27005, ISO 22301, ISO 24762, ISO 27031, PDPA, SSAE SOC 1 / 2, Data Sovereignty, Auditing, Pan Tests
- Losing Governance internally
- Forgetting importance of Domain and IT Compliance : SOX, HIPAA, PDPA, MCTS, PCI DSS etc.
- Missing BCP and DR as key backbones for their business
- Weak Marketing 

Few ways you can impress your customers:

1. Architecture
Pulling back the curtain on the architecture of the vendor’s cloud hosting platform will help you evaluate if it is right for your business. For companies looking to host business-critical applications, the cloud vendor’s underlying datacenter, network, storage, and compute infrastructure should mirror the features of a typical enterprise-class computing platform as well as offer the advanced capabilities associated with cloud computing, such as elastically scalable computing resources.
Key Datacenter Infrastructure should include:

• Fully Independent “A” and “B” Power Systems to all Physical Devices
• Fully Redundant Cooling Systems

Key Network Infrastructure should include:

• Complete Network Redundancy to the Physical Host
• Dedicated Public, Private, Backup, Administrative Networks
• SWIP and RWhois Support (i.e. ability to re-assign IPs to Customers)
• Network Layer Load Balancing
• Network Layer Intrusion Detection and Prevention
• Network Layer Firewalls
• MPLS and Virtual Private Networking Support
• Geographically Redundant DNS Service
• Edge Caching and Routing Service Options (i.e. CDN Services)

Key Storage Infrastructure should include:

• Elastically Scalable Tier0, Tier1 and Tier 2 Enterprise SAN Storage
• Fully Redundant Multi-Gbps Fibre Network
• Proven Ability to Handle High I/O Applications

Key Compute Infrastructure should include:

• Elastically Scalable CPU and Memory Resources
• Live Migration Across Physical Hosts
• Automated Recovery from Failed Physical Hosts
• Support for Oracle, SQLServer, and MySQL Clustering

Commodity cloud offerings – such as Amazon and RackSpace – lack many (if not all) of these building-blocks necessary to support enterprise-class hosting environments.Without these fundamental building blocks, the ability of the vendor’s cloud platform to adequately address core ITSM requirements and satisfy the IT requirements of medium to large organizations over the life of the project is compromised. For example, the ability to schedule application-level backups or accessing persistent data storage without impacting network performance and causing application latency is a straight forward requirement of most business-critical applications but are not supported in Amazon’s mono-network EC2 architecture.
Integrated policy and workflow engines and other advanced service management toolsets are another way enterprise cloud hosting platforms set themselves apart from their consumer-oriented counterparts. 
2. Tools
Do far more, far better, and with far less IT resources! To realize this universal goal, an enterprise-class cloud hosting platform must reinforce the integrity of ITSM best practices around application development, staging, production, and disaster recovery while streamlining the necessary IT resources required to support these best practices. This requires an advanced administrative toolset to drive efficient and effective service delivery (e.g. capacity, continuity, and availability management) and support (e.g. incident, change and release management). Key ITSM tools that should be part of an enterprise-class cloud hosting offering include:

• Web Based Control Panel/Customer Service Center
• Web Based Incident, Change, and Request Management System
• Integrated Change Auditing and Control Systems
• System Templates for Rapid re-deployment of your Application and OS Configuration
• Incremental System Snapshots with Roll Forward and Back Capability
• VM and Application Level Data Backup and Archiving
• System and Application Performance and Availability Monitoring and Reporting Tools
• Local and Remote Data Replication Tools
• Benchmark Performance and Security Testing Tools
• Service Level Agreement (SLA) Management Tools
• Web Services APIs for integration with Cloud Tools and Resources

Few cloud hosting offerings currently include all of these ITSM tools. And while it is possible to combine the services of multiple vendors to create an equivalent suite of management tools, as we saw with Amazon’s mono-network architecture, it is not always certain the underlying architecture of the vendor’s cloud will support them. 
ITSM best-practice also dictates that production systems are replicated to a disaster recovery site to ensure business continuity. Although very important, the cost of implementing and maintain this best practice is extremely high. 
3. Security
Information security is the most significant barrier CIOs see to adopting cloud hosting services. While this is a multi-faceted issue, from a technology point of view there are no inherent security risks or benefits associated with cloud computing relative to other Internet accessible computing platforms. The same principles and toolsets apply. CIOs must do the same due diligence with a cloud hosting platform as they would with their own internal IT departments or traditional IT outsourcer to make sure their Information Security Management System (ISMS) is being supported. In fact, a trustworthy enterprise-cloud vendor should enhance an organization’s ability to protect the integrity of its business information and navigate ever rougher regulatory waters. Ways this can be achieved include:
A. Requiring the cloud vendor to be certified against internationally recognized information security standards, such as SAS 70/SSAE 16, ISO 27001 and PCI. Certification against one or more of these standards is normally required of IT outsourcers to meet the regulatory requirements governing Ecommerce transactions or the management of personal and financial information.
B. Requiring the Cloud vendor to have a multi-layered information security infrastructure integrated into its cloud hosting infrastructure to protect against data intrusion, corruption, and loss. These systems include:

• Flexible Service Agreements that address Risk and Accountability
• Integrated Vulnerability and Compliancy Assessment Tools
• Integrated Change Auditing and Control Systems
• Local and Off-Site Data Backup and Archiving Services
• Integrated Network Layer Intrusion Detection and Prevention
• Integrated Network Layer Firewalls
• Integrated Virtual Private Networking (VPN)
• Support for Private MPLS Network Connections
• Strict Access Control and Acceptable Use Policies
• DDOS Mitigation Tools and Policies

C. Avoiding mass-market Cloud services. At some level, every cloud hosting service, like the Internet itself, is a shared network. As a result, what others do on that network can and will, at some point, affect your business. The risk increases exponentially if the cloud vendor is targeting the mass-market with cheap prices and automated sign ups. On this type of cloud hosting platform you can be certain that network abuse is a significant issue and customers are routinely sideswiped by Denial of Service (DoS) type attacks and blacklisted IP blocks. This is a well-publicized problem on Amazon EC2 but the problem is universal.[5] The best protection is to choose a vendor who caters to businesses and organizations conducting real business over the Internet and subscribing to sound ISMS principles.
Hosting facilities must be SSAE 16 SOC 1/2 certified with security status
5. Support Services
The majority of the cloud hosting offerings on the market are essentially self-help services. At best, organizations can subscribe to a “live” help desk service that is reactive and focused on request and incident management. This may be enough for organizations with solid web hosting experience and adequate IT staffing. However, many organizations need much more. They need a partner who is an expert in the delivery of business-critical applications over the Internet. They need someone to understand their business and application requirements in detail and work proactively through the entire IT lifecycle to achieve their goals. If this is your organization then key things to look for in a managed cloud hosting vendor, include:

• Support for all aspects of service delivery, including the application stack
• Ability to work proactively through the complete IT lifecycle
• Service Level Agreement (SLA) with end-user centric performance objectives
• Adoption of ITIL best-practices, particularly around change management
• Technical Account Managers dedicated to your account
• 24x7x365 Service Desk coverage
• Feedback collection team
• Feedback Processing Team

As one would expect, the shape of cloud hosting computing is fuzzy and changing rapidly. It is still early days and new vendors are entering the market regularly with different service philosophies, target markets, and solutions all under the same cloud hosting banner. In this environment, it is particularly important for CIOs to analyze carefully the underlying objective, architecture, tools, security, and support behind a vendor’s cloud hosting platform. Only then can they evaluate if a particular cloud can achieve the optimal balance between affordability and system availability, capacity, security, scalability, and manageability that is right for their business requirements in the short and longer term. 

Force10 VRRP

[ s4810-01 ]
interface Vlan 777 
 description [Dell_vrrp]
ip address 192.168.102.1/24
ip address 10.11.6.1/24 secondary
untagged gigabitethernet 2/24-35
tagged gigabitethernet 2/42-47
!
vrrp-group 30
  priority 150
  virtual-address 192.168.102.3
!
vrrp-group 31
  priority 150
  virtual-address 10.11.6.3
no shutdown
!
 
 
[ s4810-02]
interface Vlan 777 
 description [Dell_vrrp]
ip address 192.168.102.2/24
ip address 10.11.6.2/24 secondary
untagged gigabitethernet 2/24-35
tagged gigabitethernet 2/42-47
!
vrrp-group 30
  priority 100
  virtual-address 198.168.102.3
!
vrrp-group 31
  priority 100
  virtual-address 10.11.6.3
no shutdown
!

VLT Domain Failure Scenarios

I did some investigation tonight in my Lab running OS 9.5.0.1 with VLT and taking down the VLTI and Heartbeat Links.  Here is what I found out in case your customers ask you "What happens when…?"

Scenario 1:  All link up

Results : All links from the end-devices to the 2 VLT switches are up.

Scenario 2: VLTI link between switches goes down but Heartbeat Link is up

Results: The End-devices link(s) to the Secondary switch will go down to prevent a Loop but continue to pass traffic through the Primary switch.  If there are any Non-VLT configured interfaces they will not be effected by the VLTI link going down and continue to pass traffic normally. The Heartbeat connection is still passing the hello packet.

Scenario 3: VLTI Link up but Heartbeat link goes down

Results: Everything continues to pass traffic normally without taking any interfaces down.  The reason is that the only thing that the Heartbeat link does is pass a 56 Byte hello Packet stating "I'm here".

Scenario 4: VLTI and Heartbeat Link go down but the switches are still up

Results: Each switch within the VLT Domain will become the Primary with the End-devices connection still up passing traffic as normal.  This is call a Split Brain scenario where there is a loop within the Network as the End-Devices are passing traffic down each interface to the switches within the VLT Domain.  This is a reason to have RSTP or PVST+ (future Klamath release) configured in conjunction to VLT so in case a Split Brain Scenario happens we can have some type of backup to help prevent loops within the Network.

So the bottom line is that we want the Customers to use the OOB interface to support the Heartbeat due to the fact that if they used the VLTI and Heartbeat as the same interconnection and it went down they would cause a Split Brain scenario and cause more issues within their Network.

JOHN BECK, Enterprise Technologist | Networking

Dell | Office of Technology & Architecture | CTO

Office #: (847) 515-5910

Cell / SMS #: (847) 207-6348

Fax Number #: (847) 515-5921

E-mail: j_beck@Dell.com