Following my previous blog post on Business Continuity, I would like to continue today on the same theme, but this time moving to the concept of resilience. 

Resilience in the context of Business Continuity can be defined as a combination of the following:

Not being vulnerable
Is your infrastructure, processes and plans capable of responding to threats that you have not even thought about, the so called “Black Swan” events.

Robustness
Is your organisation able to withstand stresses, pressures, or changes in procedure or circumstance without failing?

Redundancy
Do you have fall back plans / systems / processes to keep the business running?  These redundant options may require the business to chop and change in the event of the disaster …

Adaptability and Agility
Is the organisation agile enough to adjust to the changes imposed by the above?

Resilience is a property of not an activity.  That is to say it is what your organisation is that makes it resilient, not what it does.  Therefore resilience is a function of the People, Processes, Technology and Partners.  Moreover, resilience is determined by what your organisation is capable of doing, rather than the number of plans and documents you have.  Practice and rehearsals are much valuable to developing resilience than the plan artefacts, which reinforces the points made about BCP above. Resilience can be developed in an organisation, but it is difficult to measure.  However the level of resilience in an organisation can be observed during and after an incident. 

As with BCP, building resilience is a journey, not a destination.  As your organisation continues to evolve, so must the business continue to think about ways to ensure it remains resilient.

For further information on resilience, business continuity and disaster recovery, I would recommend reading Ken Simpson’s blog here.

The topic of the last PicNet User Group meeting was on Business Continuity and Resilience.  The discussion in this meeting revolved around the importance of business involvement and eventual business ownership in a successful BCP project and how to develop resilience in an organisation.  It is from these discussions I would like to share some of the thoughts on this subject.

All too often the terms Disaster Recovery and Business Continuity are (incorrectly) considered to describe the same thing, that is, the recovery of IT system after a disastrous event.  This has lead to a trend for the responsibility of creating and maintaining a Business Continuity Plan to be left to IT.  However, Disaster Recovery (DR) should not be confused with Business Continuity, for which DR is only a subset dealing primary with IT.  Business Continuity Planning (BCP) must include areas of business, not just the IT functions.

The experience of producing DR Plans puts an organisation’s IT department in good a position to enable and facilitate the BCP process.  However for BCP to have any chance of success, it is imperative that all functional parts of the business take an active part in the development of the Business Continuity plan.  Furthermore, long term success is dependent on the business taking ownership of the plan once it has been produced and dedicate the required resources in plan maintenance to keep it up to date.  One our customer has had direct experience of this.  Several years ago, this customer’s IT department undertook a project to produce a Business Continuity Plan for the for the wider business.  However, the BCP process had not been sufficiently explained to the business, which resulted in a lack of resources from the various functional departments.  Without sufficient business involvement, the project inevitably failed.  In 2009, the IT department once again took the lead in producing a Business Continuity plan.  However this time around, the process and the requirements were first explained to the Managing Director and the various departmental directors.  It was made clear that for BCP to succeed, the business as a whole must be involved and resources made available.  Once the plan had been produced, the business must take ownership of the plan and ensure it is maintained and kept up to date.  With the boarder business onboard, this second BCP project was successful.

When it comes to BCP, often it is the process rather than the end artefact that is of the most value to an organisation.  The process of planning for business continuity stimulates thinking on processes, organisation, suppliers, customers etc and ways to keep the business operating if the proverbial “sh*t” was to “hit the fan”.  The process of testing the BC plan either reinforces these ideas or (if it is found that the processes detailed in the plan are no viable) forces new solutions to these problems to be found.  This is also true for the maintenance of the plan.  The maintenance process allows the business to increase their awareness of the BC plan and to re-examine their business to verify that the details of the plan reflect the needs of the current business.

For further reading, I would recommend reading Ken Simpson’s blog here. 

More to follow in my next blog entry on Resilience.

After working with the Riverbed Steelhead Appliances for a number of years now I thought I would share some of the additional performance tweaks you can make beyond the out the box configuration without the need to contact Riverbed Technical Support. Some of the tweaks here require that you have the latest version of the Steelhead Operating System RiOS 6.0 which was released at the end of 2009. For those who have never heard of Riverbed they are the current market leader in WAN optimization products.

PicNet is a reseller of all Riverbed Products and Services.

1. Enable High Speed TCP (Configure -> Optimization -> Performance)

This is useful if you have a high bandwidth link with relatively high latency. To know whether you need to enable high speed TCP you need to calculate if your Bandwidth Delay Product (BDP) is greater than 256KB. To do this calculation:

BDP (in Kb) = 2 x Bandwidth (in bits per sec) x delay (in ms) / 8 (bits per byte)

2. Selecting a Data Store Segment Replacement Policy (Configure -> Optimization -> Performance)

You have two options here for the data store policy either First-In, First-Out (FIFO) or Least Recently Used (LRU). All Steelhead Appliances prior to RiOS 6.0 had the default storage policy of FIFO but in RiOS 6.0 they have changed the default to LRU. Once you perform a RiOS upgrade to 6.0 your storage policy will change to LRU. This is a smart move in my opinion as it will allow for potentially more data reduction.

3. Optimizing the Data Store for High-Throughput Environments (Configure -> Optimization -> Performance)

This setting may come in handy if you have high throughput workloads such as SAN replication for DR etc between two Steelhead appliances. To maintain consistent levels of performance, Riverbed recommends using separate Steelhead appliances for DR workloads than for optimization of other traffic. You basically have three options to choose from:

• Default – which provides the high level of data reduction (default setting) 
• SDR-Adaptive – dynamically adjusts data reduction vs data throughput based on loads being placed on the Steelhead appliances. 
• SDR-M – Performs optimization in memory only so there is no latency burden by the physical disks in the data store. In this mode the datastore is not used at all any latency/traffic improvement are done in memory only. This mode maybe useful for SAN replication.

4. Configure CIFS Optimization (Configure -> Optimization -> CIFS)

There may be some additional tweaks you can make here to your environment to enable the acceleration of Windows File sharing traffic.

Optimize Connections with Security Signatures (aka SMB signing)
With older versions of RiOS they were unable to fully optimize CIFS traffic when Windows was configured to sign the traffic. This was especially true for File Shares hosted on Domain Controllers where this was the default setting. You either had to disable SMB signing in your Windows Environment or accept the fact that you would not get the full optimisation benefits. Now you have the option of making the Steelhead appliances members of the Windows domain environment to allow full optimisation without jeopardising security. To enable this setting you must make the Steelhead appliances a member of the domain first before enabling SMB signing otherwise you will get a warning. Secondly you need to choose the SMB mode either Transparent or Delegation with the former being far simpler. Delegation mode only needs to be selected if you had SMB signing configured in RiOS 5.5x and requires additional configuration not mentioned here.

Enable Dynamic Write Throttling
replaces the default static buffer scheme. When enabled it will dynamically adjust writes if there is a backlog of write messages on the server-side device.

Enable SMBv1 Backward Compatibility
For a performance increase between clients that are Windows Vista and Windows Server 2008 it forces communication from SMBv2 to SMBv1 which allows for higher performance. At this stage I’m not sure if this also affects Windows 7 and Windows 2008 R2 where they use SMBv2.1 which is supposed to enhance CIFS performance further.

Enable Print Optimization
Improves centralized print traffic performance. For example, when the print server is located in the data center and the printer is located in the branch office, enabling this option speeds the transfer of a print job spooled across the WAN to the server and back again to the printer. Note: This feature does not improve optimization for a Windows Vista client printing over a Windows 2008 server, because this client and server pair uses a different print protocol.

5. Configure MAPI Optimization (Configure -> Optimization -> MAPI)

Exchange 2007 Optimization
In older versions of RiOS prior to 5.5x Exchange 2007 acceleration was not supported and then when it was supported it did not support Encrypted traffic by Outlook and hence encryption had to be disabled in Outlook 2007 clients to gain optimization benefits. Thankfully this has been rectified and is no longer an issue in RiOS 6.0. You should now enable this optimization if you use Outlook 2007 in your environment.

Enable Transparent Population 
This setting basically allows email for users located in branch sites that do not have a local Exchange Server to continue to pull down their mail even when they have Outlook closed. That way when they open their Outlook they receive any new email at LAN like speeds. This maybe particularly useful to curb the surge of network utilisation in the morning when remote users logon to the network and fire up Outlook.

6. Configure Citrix ICA optimization (Configure -> Optimization -> Citrix ICA)

With the introduction of RiOS 6.0 brings about new Citrix Optimisation which allows optimization of the ICA protocol bandwidth as well as being able to provide QoS shaping and priority.

To enable this setting is quite straight forward:

• Specify the ICA ports (usually 1494 and 2598) and exclude them from the Inter active Ports label. 
• Enable the Citrix ICA Optimization 
• Optionally enable QoS priorities.

7. Configure HTTP Optimization (Configure -> Optimization -> HTTP)

There are a few minor tweaks you can do over the default configuration here and that is around the areas of:

• URL Learning (enabled by default)
• Parse / Prefetch
• Object Prefetch Table

URL Learning - The Steelhead appliance learns associations between a base request and a follow-on request. This feature is most effective for Web applications with large amounts of static content, for example, images, style sheets, and so forth.

Parse / Prefetch
Parse and Prefetch essentially reads a page, finds HTML tags that it recognizes as containing a prefetchable object, and sends out prefetch requests for those objects. Typically, a client would need to request the base page, parse it, and then send out requests for each of these objects. This still occurs, but with Parse and Prefetch the Steelhead appliance has quietly perused the page before the client receives it and has already sent out the requests. This allows it to serve the objects as soon as the client requests them, rather than forcing the client to wait on a slow WAN link.

Object Prefetch Table
The Steelhead appliance stores object prefetches from HTTP GET requests for cascading style sheets, static images, and JavaScript files. This helps the client-side Steelhead appliance respond to If-Modified-Since (IMS) requests and regular requests from the client, thus cutting back on round trips across the WAN. This feature is useful for applications that use a lot of cacheable content.

To enable and test these 2 settings you may want to set up another server subnet (say that your internal web servers reside in) and apply one or both of these settings and test the results.

I have provided a comparison here from Riverbed of the different HTTP optimization features.

This is a great article put up by ITnews.com.au for any IT Manager out there:

http://www.itnews.com.au/News/166078,five-tips-to-kit-out-your-data-centre-on-the-cheap.aspx/0

Hello and welcome to my first new blog post of the decade...aka 2010!

I recently underwent a research of the various big cloud providers as part of a new project / product inside PicNet and I thought this would be a good place to share with you some of my thoughts on the differences between the big 3 cloud providers: Amazon, Microsoft & Google and some other competitors such as GoGrid, The Rackspace Cloud and Melbourne IT.

I will also touch on VMware’s Cloud Platform (vCloud Express) which is available from a number of providers.

With so much noise being created around cloud computing it will gain even more traction and attention in 2010.

So without further ado here are my findings presented in a nice table format:

Prices correct as of Jan 2010 (click to expand)

Conclusions

• It seems that with most providers that storage and data charges are the same amongst providers

• Compute power pricing differs a fair amount because it is difficult to get equal compute power from multiple providers

• Amazon has the most established, stable platform offering the widest array of support for different technologies - hence is current king of cloud.
  

• The only offering in Australia at present is from Melbourne IT which is a limited beta of VMware's vCloud Express platform. I expect more cloud providers to start popping up that run this as the underlying platform.

• Google is the only provider to allow you to get going for free for small apps, however its u

  se in the enterprise is an unknown at this stage.

• Security and Privacy of Cloud Providers is also a bit sketchy at this stage with not a whole lot of documentation available on how your data is protected.

I stumbled across this article this morning and found it quite an interesting read. The article describes why many organisations struggle with IT due to a lack of standard definitions for terminology and practices.

Click here:

Microsoft has recently released the new version of Exchange Mail Server 2010. As with each new version of Exchange new features are added and it's very nice to see that the focus on new features around retention, governance & compliance and the ability to get rid of PST's in the organisation! Disparte PST's in the Corporate Environment have been a nightmare to manage for sys admins as they are often stored on the local computer and are not backed up at all and the fact they are stored on portable devices makes the information prone to theft and not retrievable. Previously such tools were only possible through 3rd-party vendors - while the Microsoft tools may not be as comprehensive we will need to wait and see.

Another offering that will be coming in 2010 is Hosted Exchange Online where Exchange will run out in the cloud by Microsoft. It will be interesting to see what kind and size of companies would be keen on sending their messaging infrastructure out to the cloud.

PicNet being a Microsoft Gold Partner already has access to this new software and I will be giving this a roadtest in a virtual environment in the coming weeks to see how this new messaging platform performs as well as the new features

The majority of projects that I have managed have been software development projects. At PicNet we use an agile software development methodology.

We like the Agile approach because it allows us to get feedback from the business on the system in development at a very early stage, thus enabling us to fix business logic and usability issues early in the project lifecycle.

This reduces the possibility of costly late functional requirements changes and also facilitates more business involvement in the development of the system, which fosters a culture of system ownership in the business. Ever since I start project management I have been curious as to how software development methods such as Agile fit with into the larger picture of project management, in particular how do they relate to methodologies such as PRINCE2?

I recently attended a PRINCE2 course and it was established that software development methods such as Agile, RAD, waterfall etc fit into the delivery layer of the PRINCE2 method. As PRINCE2 is a universal project management methodology, it provides great detail for directing and managing a project, but is deliberate silent on the method for actual product delivery. Software development methods such as Agile compliment overarching project management methodologies such as PRINCE2 and in fact the DSDM-Atern agile software development methodology is fully compatible with PRINCE2. A white paper on DSDM-Atern and PRINCE2 can be found here.

We are still in the early stages of incorporating our agile software development approach with the PRINCE2 method, and I would be interested in hearing from people who have gone through the process already.

I've just come back from spending 2 days at the biggest virtualization forum in the ANZ region for 2009.......the VMware vForum!

I must say that the event was huge with over 2500 attendees and over 60 technology vendors with their products, solutions and services on offer all circling around virtualization!

Cloud Computing being the new buzz was high point of discussion throughout the Forum. It is quite outstanding to see VMware as a market leader making leaps and bounds in this area and making it a reality. Here in ANZ we got to hear about how Telstra, Optus and MelbourneIT have already built or building their own public cloud offerings.

Now you might be thinking what exactly is cloud computing? Well the definition that seemed to come out of the forum is that:

Cloud Computing = IT as a Service (ITaaS)

in other words just like your telephone or electricity you only pay for what you use and IT resources should be the same.

So what else did we learn? Well got to learn a fair bit about VMware's latest offerings vSphere 4 and all the new features and how you can leverage these in the organisation. Features such as automating much more of the remedial tasks of provisioning VM'sm, mananging them etc.

One cool feature is the Dynamic Resource Scheduler in conjuction with Dynamic Power Management that move workloads to other servers afterhours say and shutdown those servers completely saving power.

I'm definetly keen on trying out VMware's offerings myself, for PicNet and our customers.

Cloud computing and green solutions high on the list for next year!

1. Cloud computing
2. Advanced analytics
3. Client computing
4. IT for green
5. Reshaping the datacentre
6. Social computing
7. Security – activity monitoring
8. Flash memory
9. Virtualisation for availability
10. Mobile applications

Click here for more detail: