SIP Trunking with your MPLS network

January 15, 2010 by Leave a Comment

A google search (June 2009)  on a ‘SIP Trunking’ returns ~500k references. 

From Wikipedia (replace ‘connection’ with Trunk)
 A SIP (Session Initiation Protocol) connection is a service offered by many ITSP (Internet Telephony Service Providers) that connects a company’s PBX to the existing telephone system infrastructure (PSTN) via Internet using the SIP VoIP standard. 

This probably is the most common understanding of SIP Trunking – that being, a SIP-based interconnection from a SIP-based IP-PBX to a SIP Service Provider  which facilitate communications to and from the PSTN.  Within this context, there’s ample evidence to the value of SIP Trunking.  The first, and probably most obvious advantage is a reduction of costs by leveraging IP as the mode of communications between the Enterprise and the Service Provider. Historically, businesses had two distinct and separate network infrastructures – one for data, and another for voice.  Maintaining one is easier (but a little bit more complicated at times) than two.  If done correctly, it’s definitely cheaper.  This point pertains to the LAN as well as the WAN.  

Now, if the PSTN were the sole interconnection of interest for SIP Trunking, owning and managing SIP Trunking from the Service Provider POV would be easier. (not easy, but easier) BUT, here things are changing – as more and more businesses (Enterprises and/or Carriers) Transition to SIP, the burden of ensuring that all the SIP Trunks interact successfully with one another falls on the Service Provider. All SIP networks are not built the same.  There are numerous differences between Enterprises and Services Providers alike.  

Possibly this analogy will help with the understanding.  Say each SIP-based network, represents a different language, French, Chinese, Spanish, etc.  In order for everyone to communicate with one another, everyone would need to be able to speak and understand everyone else’s native language.  There must be a common denominator.  Without a common language, then translation services are necessary to contract out to facilitate the communications.  The PSTN has historically served as the “common denominator” in the world of IP Telephony integration.  The problem stems from the cost of using the PSTN in this manner, both tangible and intangible.  First the tangible costs.  It’s more expensive to use the PSTN for this reason.  Secondly, things get lost in translation.  If the PSTN is used as the common denominator, then features other than standard voice (video, HD voice) are not possible, and there’s oftentimes service degradation from multiple encoding/decoding of the voice.

The SIP Trunking Service Providers (at least the good ones) ensure that the necessary translation services required to enable communications do NOT limit the capabilities of SIP, and still enable the end-users to realize the savings of SIP Trunking.  This is the true value of the SIP Trunking Service Provider.  The Enterprise only needs to ensure that the communications interconnection (SIP Trunk) between them and the SIP Trunking Service Provider are compatible, and the rest can be assumed.  This is something that is too often overlooked in the world of SIP Trunking.

Another little know fact is that with international calls, many of the carriers are already using SIP, even if the caller is not.  If you start with SIP from your network , you eliminate one of the encoding/decoding steps and end up with better voice quality.

Remember, I am talking about SIP trunks using an MPLS network…NOT the internet.  This eliminates the lack of quality control that come to mind for most people when they think of SIP (think Vonage).

Filed Under: Network FAQs, New MPLS Implementations, Notes About Implementations, SIP Trunks, Uncategorized Tagged With: , , , ,

Making a router choice: Cisco 1841 or Cisco 2811

November 13, 2008 by Leave a Comment

When you are deciding on an MPLS network carrier, an important thing to consider is the router that will be provided, assuming that you want fully managed services.  This is a subject that many people are unclear on. 

Many carriers will provide a Cisco 1841 in their quotation.  The primary reason being cost.  Some carriers will quote a Cisco 2811.  While the cost is a bit higher, so in the expandability and performance.  While this blog is not intended to be a technical hardware discussion, the general subject of MPLS networks warrants a brief message about this topic.

Cisco 1841:

  1. List: $995 (depending on configuration)
  2. Memory (Std/Max): F: 32/128 and D:128/384
  3. LAN: 1 E/FE
  4. WAN Card Slots: 2 (data only)
  5. Network Module Slots: none
  6. Advanced Integration Module Slots: 1
  7. Digital Signal Processing Slots: none
  8. USB Ports: 1 (for USB Flash configuration)
  9. Form Factor: shelf
  10. Power supply redundancy: none
  11. Voice support: none
  12. Fast/CEF Switching: 38.40Mbps with 64 byte packets

Cisco 2811

  1. List: $1750 (depending on configuration)
  2. Memory (Std/Max): F: 64/256 and D:256/768
  3. LAN: 2 E/FE
  4. WAN Card Slots: 4 HWIC/VWIC/WIC/VIC (Hardware WAN Interface Card/Voice WIC/WIC/Voice Interface Card)
  5. Network Module Slots: 1
  6. Advanced Integration Module Slots: 2
  7. Digital Signal Processing Slots: 2
  8. USB Ports: 2 (for USB Flash configuration)
  9. Form Factor: rack
  10. Power supply redundancy: RPS-675 (optional power supply)
  11. Voice support: DSP/NM
  12. Maximum CCUE phones:  36
  13. Max SRST Phones: 36
  14. Fast/CEF Switching: 61.44 Mbps with 64 byte packets

To summarize:

  1. The 1841 is suitable for a single T1 or E1.  If you plan on using VoIP or growing to NxT1 or NxE1, select a better router.
  2. The 2811 is suitable for two T1s or E1s and supports voice.  It’s little brother the 2801 has less memory and less expansion slots, with a base configuration for a single T1 or E1.
  3. If your growth requirements will be greater, get the right router from the start to allow your upgrades to be painless and avoid any unexpected capital purchases.
  4. There are other router manufacturers than Cisco, so consider your options in full.
  5. Carrier provided routers and management will cost you more than leasing and managing the routers yourself, providing you have the resources to do it yourself.  Managed routers eliminate the management headache.
Filed Under: Notes About Implementations, Quotation Thoughts, Uncategorized Tagged With: , , , , , , ,

Important facts pertaining to VoIP and Wide Area Networking

September 25, 2008 by Leave a Comment

Quality of Service or QoS is the quality of a call over a network. It also refers to the ability to prioritize certain types of traffic on an IP network. In the case of VoIP, this typically means prioritizing voice traffic at a higher level than other forms of traffic such as data so that voice traffic will not be delayed or dropped.  An MPLS network will allow you to prioritize this data ahead of all other data traffic in order to maintain quality.

Latency causes delays in packet delivery. Physical distance, the number of router hops, encryption, and voice/data conversion all impact latency. Users begin noticing latency as a service level issue when roundtrip latency is greater than 250 milliseconds (ms). The International Telecommunications Union recommends that latency never exceed 300 ms round-trip.  Over long distances, i.e. from the USA to Asia, the shortest path circuit and least number of hops can make the difference between satisfactory and unsatisfactory voice communications.

Jitter occurs when voice packets are sent and received with timing variations. Jitter is effectively a variation of packet delay where delays actually impact the quality of the conversation. Think of jitter as variable delays in packet delivery. Participants will notice delays in the conversations impacted by jitter. As a result, many service providers now account for maximum jitter levels.

Packet Loss takes place when packets are dropped.  This can be due to a variety of factors and is very common when using VoIP over the internet, with or without a VPN tunnel. It usually shows up as dropped conversations or “tinny” sounds. Packet loss should never exceed 1% and most service providers guarantee service levels with .5% or less packet loss. Packet loss of 1% translates into one voice clip or skip every three minutes, while packet loss of .25% translates into one error every 53 minutes.  When latency is high, packet loss is often high, as well.

Prioritizing VoIP traffic over the network at Layers 2 and 3 yields latency and jitter improvements. Policy based network management, bandwidth reservation, Type of Service, Class of Service, and Multi-Protocol Label Switching (MPLS) are all widely used techniques for prioritizing VoIP traffic at Layers 2 and 3.  VPLS networks will provide the highest level of quality available, where appropriate.

Filed Under: Industry Surveys, Notes About Implementations, Uncategorized, VPLS Tagged With: , , , , , , , ,